Our first speaker today is my co-moderator, Dr. Carolyn Quincy. Perfect. So I'm going to kick off our session just with a brief couple of comments about quality of life and outcomes. Thank you so much for attending today's tumor session. We'll have our talks here and then a beverage break and then we'll continue with our talk. So I just wanted to say that quality of life and reported outcomes is incredibly complex. There's so many different ways that we measure it. As someone who treats children, sort of have the built-in systems that really tells us a lot how children's executive function is doing. That's really a challenge when it comes to treating our adult patients and figuring out how their executive function and things are going. So I think it's a real challenge. It's one that we should all strive to measure in order to improve it and I think one that is definitely worth investing in. So I'm going to turn it over to our first talk which is going to be Eric Oberman. Oh, there he goes. Does this work? The mouse it is. Yeah. Awesome. First, I wanted to thank the tumor section, especially Dr. Germano, for inviting me to speak today about artificial intelligence and brain tumors and the path forwards to get to really offer sort of my opinion on sort of where we are and where we're headed. I do have some financial disclosures not relevant to this talk. I also have a disclosure from Master Yoda that it's hard to predict the future, so there's a good chance that everything I say today could be wrong. But I think there's also the intriguing possibility that at least some of it might be right. And so where are we with AI and brain tumors? I think 2023, the bulk of what we do with AI and brain tumors, largely consists of training models for classification on imaging data or tabular data, really to predict diagnosis and prognosis. And this is in many ways fairly similar to what we've been doing for the past 10 years, 20 years in this area. There's definitely a lot more machine learning these days. There's more deep learning these days. But in some ways, in 2023, looking at the field as a computer scientist, I feel like the big divergence is where we are with AI and brain tumors from AI and computer science, which, as many of you have probably seen in the news, and maybe even in your personal lives if you've interacted with these tools, that generative AI is really completely upending a lot of how we interact with society, certainly how we think about AI and computer science. So I think there's an increasing gap then, perhaps, in terms of where we are as a field in AI research and where we are in brain tumors. And that's what my talk is going to address, really, that gap, which I see as being the biggest barrier to the path forwards. These are the basic topics I'm going to touch on. Really, if there's one theme to my talk, it's technical debt, which I'm going to define that term in a second. But it's also really starting to bridge the gap, as I see it, is moving from simply writing papers about AI and brain tumors to using AI for brain tumors. This is comical, but also, in my mind, this is the attitude that I think we need to start having, of that, look, we have a cool algorithm and it works. Let's not just write about it, let's use it. And the barrier to that, really, in my opinion, really, it's technical debt. And so this is a term from software engineering, and it refers to the fact that lots of times in engineering, when we push systems out into production, when we start using them, we'll make design choices to get them functional that maybe limit our ability in the future to integrate them or to update them. And I think, certainly, over the past decade, the push to roll out electronic health records has incurred for all of our health systems and our practices a substantial amount of technical debt. Modern EHRs, beyond the fact that they're not even user-friendly, I doubt anyone here likes their EHR, they're also incredibly hard to integrate with and incredibly hard to extend. And so the technical debt that our health systems have really, in a sense, imposed on us with EHRs, in my mind, is one of the biggest gaps in the barriers to the path forward for using AI in brain tumors. But I do think, also, we're starting to see research groups within the tumor section, within our field, starting to cross this gap. Certainly, hospital IT departments, I think, are beginning to be aware of this. Fundamentally, IT groups don't hire software engineers, and they don't do this kind of work, but they're starting to. But I think even more encouraging is that, really, the push to move forwards is coming from our departments. I would argue that the hottest hire in a non-faculty hire in a modern neurosurgery department is an engineer, a full-time software engineer, to really work on questions of integration. And perhaps alongside this push of starting to onboard engineers into neurosurgical departments, I think we're increasingly seeing a fluency of neurosurgical faculty and neurosurgical residents with enough competency in programming and software engineering to start building software that not only uses some interesting algorithm and writes some interesting paper, but actually starts to get used. One of my favorite projects is the DeepMets initiative. This one is probably not as well known to this group. This is out of Taiwan AI Labs in Taiwan, and it's a multi-center study. It really runs across about 20 different hospitals where they've aggregated their data on brain metastases, and they have a software program which does things that we've seen in the literature, maybe some of us have even published on, of automatically detecting metastases. What is so cool about this, though, is that they have not just written a paper on this, but this is integrated into their hospitals, into their treatment planning workflow. If you have a patient with brain metastases at the Taiwan General Hospital, you can get an MRI scan, push it to the software, it'll automatically identify the brain metastases, and then push that to GammaPlan. And so this is software that's not just detecting brain metastases in the academic literature, but it's actually being used, and it can actually be studied for how does this really impact our practices, how does this really impact our patients. This is another more modest project. This is a recent project of ours out of NYU. If you have a glioblastoma at NYU, the volume of that tumor is not going to be measured by your neuroradiologist. It'll actually be done algorithmically. The MRI scan will get pushed to an algorithm, which will automatically segment the tumor, calculate the volume, and then write that in the report. And I think initiatives like this, we're not only writing papers about, we can automatically calculate the volume of a glioblastoma, which is in some sense not interesting, but we can automatically calculate it and write it into the radiology report, where any physician can just read it and see this is what the tumor volume is at this time. This is the path forwards. Similarly, I want to note out of the project of the Department of University of Pennsylvania's neuroradiology group is a fantastic computer science team. They have the software Gandalf, interestingly named, but really looking at trying to tackle the gap of the preprocessing of our data for all of these algorithmic works, where a lot of the difficulty of just assessing the literature are the differences in how we preprocess the data, how we source it. And open sourcing the code so we can all have a uniform preprocessing pipeline and ultimately build algorithms then that are easy to compare against one another and hopefully deploy, again, the path forwards. So for me, 2023 is starting to get a taste of this path forwards in terms of deployment. I'm actually getting to use some of these predictive tools in our practices and not just think about them and write about them. I certainly think that generative AI that I mentioned earlier is a potential big game changer. This was generated by an AI, by the way. It's a picture of the Swiss Alps. You probably wouldn't know just looking at it. We're starting to see this penetrate the tumor space. There's a recent work, it's not peer-reviewed yet, out of the Mayo Clinic, looking at using the exact same technology that drew that picture of the Swiss Alps with, in a sense, drawing MRI scans. In the top left, this is a patient with a cystic glioblastoma. Drawn over and using the exact same technology, here's the same MRI scan with that tumor removed, perhaps post-operatively resected. I think there's a lot of potential for these generative techniques such as this. In my mind, the biggest barrier to using generative AI in the brain tumor space is really thinking of how are we going to use this. What are the questions we want to ask and deploy this technology towards, which is really up to this audience. Certainly we've seen, even I think last week, Microsoft announced that GPT-4 is going to be integrated in Epic over the summer. But how does that penetrate our neurosurgical practices? I think it's really up to us. I think another part of the path forward is open data. Here I have an example. This is the SuperGLUE dataset. It's a natural language processing dataset out of NYU. It's non-medical. But I wanted to highlight the fact that for the top 13 teams on this dataset, it's 13 different research groups and it's really spread across the world. I think one of the unique things that we do well in computer science is we open source datasets which facilitate international collaborations, which we don't really do in medicine right now. There's a few good reasons for that. There's certainly a lot of bureaucracy with open sourcing data. There's definitely concerns over privacy. It's actually very challenging to de-identify large medical datasets. Certainly there's also a cultural element to this. I think in medicine we're very accustomed to realizing there's a lot of value to our data and we tend to hoard it as investigators. But I think especially in the brain tumor space, this poses a big barrier to the use of AI. Using AI techniques, which typically require tens of thousands, hundreds of thousands, nowadays we're seeing AI imaging datasets with trillions of images. It's inconceivable to me that we will have a full use of AI for the treatment of brain tumors unless we start sharing data. Certainly we've taken this approach at NYU. This is a talk that's ultimately going to happen tomorrow, but we open sourced recently all of our brain metastasis patients. You can go online and if someone's had a brain met that's received radio surgery at NYU, you can download the scan, the segmentations, and the clinical course. I think similarly, out of UPenn, the BRATS initiative, which has been going on for almost the past decade now, I think everyone in the room who's done neuroradiology research with brain tumors has probably encountered the BRATS dataset. I think it can be transformative when groups share data like this in terms of using AI for brain tumors. And when we can't or don't want to share data openly, I think there's still ample room for collaboration. There's a recent paper by Holland et al. in Nature Medicine on using convolutional neural networks and stimulated Raman histology to predict gene mutations in gliomas. The paper's a fantastic paper. I highly recommend reading it. The AI techniques are well done, but to me the most impressive thing about this study is that it involves a multinational, multi-institutional collaboration of about 20 different sites to bring together hundreds of patients with gliomas for the purposes of training this AI and then validating it across multiple sites. And I certainly think that collaborations like this are the path forwards. And beyond sharing data, certainly there's room for sharing our code too. There's a lot of redundancy in AI programming. I'm sure we all use the same infrastructure when we work with neuroradiology data, maybe even in some cases the tabular data. Certainly on the publishing end, I think there's room for encouraging code sharing. We're definitely planning this for the journal, Neurosurgery. But I think increasingly realizing that certainly the value is in our data sets and our ideas and that the more we share code as a community, it'll reduce redundant work and also lead to more reproducible science. So the path forwards, I think it's going to be solving this technical debt, which ultimately we're going to overcome with sharing code and with working together, perhaps even with hiring engineers or educating ourselves to do more software engineering. And ultimately this is going to lead to actually using the AI that we talk and write about, ultimately to advance our research and our patient care. I'd like to thank my fantastic team, the Dawg Auto, and I'll be available after this session for questions. Thank you. Thank you. Now we'll hear from Dr. Newman. Perfect. All right. So I'm Chris Newman. I'm from Sloan Kettering. And I'm going to talk a little bit about the present and the now and financial toxicity and spine oncology, something we don't really talk a lot about. But I think from a quality standpoint, we actually should start talking a lot more about no relevant disclosures. So we can start by just saying, what is financial toxicity? Because it's a term that some people are familiar with, some people aren't. But really it's just getting at objectively and subjectively trying to quantify the amount of financial burden our patients are feeling from their care, and in particular from their cancer care. And I really think about this as boiling down to monetary and psychological impacts of this. And when you look at the examples that come up easily, a lot of them will start dealing with things like, are patients rationing their pills because they can't afford their next one? Are they cutting back on leisure activities that they actually enjoy? Are they dipping into their savings? Are they skipping out on treatment because they can't afford to make them and they're otherwise under-treating their cancer and actually having inferior outcomes? And when we look at what are the risk factors for this happening, it's all the things that we'd find to be naturally intuitive. Inadequate lack of health care. If you are under-insured or not insured at all, you're bearing the brunt of this out of your own pocket, and that winds up being a problem, especially in younger age people where they oftentimes have high deductibles for their insurance and wind up actually paying large out-of-pocket expenses and not being able to afford their care. You have a longer treatment course, or if you're doing really well on surviving on a new medication, you're at a higher risk for this because the monthly cost, and I'll show that in a slide later, can be so high in terms of the out-of-pocket expenses that you're facing. And obvious things like if you're the primary earner for your family. If you're no longer getting an income or a paycheck from your job because you're getting cancer care, you're at a higher risk for this. But all these things kind of intuitively make sense. So if we know what it is and we know what the risk factors are, we should talk about how do we actually measure it. And measuring financial toxicity actually is interesting because there's not a single way that people do it. You can kind of break it down into objective and subjective ways of doing it like we talked about before. On the objective end, you're really trying to quantify the amount of financial burden that someone has. So are they utilizing your hospital's financial assistance and services programs? Are they going through bankruptcy? Are they having bills go to collections? That's really an easily quantifiable number. The other half of that that I would say is equally important is actually understanding the subjective part of it, of what they're experiencing. And a lot of that boils down to the questionnaires and surveys that people will have. You can see one of those surveys on the right side of the screen there, which is the In-Charge Financial Distress and Well-Being Scale. And there's other ones like this, like the cost survey and other ones. But what they really try to get to is the subjective feel that the patient has. It's not just are your bills going to collections. It's do you feel like you can afford to pay things? Do you feel the stress of this? And this gets to the quality of life part that our cancer patients actually are constantly dealing with of do they actually have the means to be able to get to the next month, to afford the things that they need to, or to enjoy life while we're prolonging life with our treatments. And so the question that always will come up is why is this important? Why should we care about it? And the big issue here is cancer costs aren't going down. They are expected to go up by 34%, to hit $245 billion by 2030. And the interesting part is that if you look at the 2020 number, where it's $200 billion, about 10% of that, or $21 billion roughly, is what the patients are actually paying themselves. So it's a big cost to the healthcare system, but it's also a big cost to patients. And if you think about we have better drugs, we have newer drugs, when you look at that bottom right portion of the slide, those are the monthly costs that insurance is hopefully paying most of, but patients are seeing a fraction of that. And with deductibles going up and the fact that 42% of people who are newly diagnosed with cancer will go through bankruptcy or deplete their savings within two years, it's a problem that we should start addressing and really start quantifying of how much do we have this in our population. Because if you start looking at the broader population of other cancer types, what you'll find is a wide range. You can find people who are doing kind of quantitative monetary estimates, saying somewhere between 20% and 50%. If you start acting subjective things with these questionnaires, it can be as high as 73% in certain cohorts. And all of this has been shown in numerous other cancer literature and other cancer types to be correlated with quality of life, symptom burden, and survival. The higher your financial toxicity is, the lower your quality of life, and the less likely you are to survive because you're not getting your treatments. And if you are getting your treatments, you're spending all of your time stressed out about it, not spending time with your family in a quality way, and just having a poor overall quality of life. And so if we look to neurosurgery and say, what are we doing in the neurosurgery field, it really is this. This is the primary paper that's out there from our field. And what it is is it's a paper out of Stanford. They looked at 93 patients who got radiation treatment alone for their brain and spine mets, and they asked them the survey question. This is called the personal financial wellness scale. It's the same as the in charge that I showed you before, some slight modifications. But what they saw was that if they prospectively asked people the question, given the questionnaire, do you experience difficulty with paying for things? If you had an emergency bill, could you pay that $1,000 bill if it happened the next month? About a quarter of all patients met criteria for financial toxicity. And so at Sloan, what we tried to do was to say, okay, we want to look at this in our patient population, but where do we start? And so we started on the more objective side to actually quantitate, do we have a problem here? How high is this problem? Is it worth further investigating? And what we did was we looked at kind of the different metrics that we can actually pull from our data set, or from our hospital EMR. And so we looked to see if bills went to collections. And not just one bill, but two separate bills going to collections. So it wasn't you missed the bill in the mail and it went to collections, but it was a repetitive pattern. We looked at people who declared bankruptcy, people who had bills that went to settlement. We looked at people who entered into payment plans to try to be able to pay off these things, people who utilized the co-pay assistance programs that we have, noting that those are mostly for chemotherapy and other kind of pill-based medications or IV infusions. Looking at utilization of the financial assistance program, which is a bigger, more holistic, it covers a lot more things related to not just the chemotherapy, but other aspects of your medical care. And then also looking at our social work philanthropy fund, which covers all the non-medical things in terms of getting your ride to and from your treatments, or even helping you just pay your rent. And then we looked at did the patient actually just express a concern to someone? Because in our hospital, if you express a concern to your physician or to someone, I'm not sure I think I need some help, it triggers a social work consultation that then gets flagged in the charts. We're able to pull all these things out to try to understand how bad is this problem in kind of our population. And it's an interesting question because our population is a very highly resourced population when you think about the tri-state area, where you have New York, New Jersey, and Connecticut that are top 10 wealthiest states in the country. And so what we did was we actually did a retrospective review where we looked at all patients who had a spine oncology diagnosis who underwent surgical treatment, with or without radiation afterwards, between 2018 and 2021. So there's some pandemic years in there. Everybody was an adult. We didn't delve into the pediatric population for this subset. That's kind of another beast that we've tackled because it's a very different thing dealing with CHIP and other services. And then we looked at kind of treating the surgical date as the index date. And we looked two years prior and two years afterwards to see what were the rates of financial toxicity. Was it coming more in the pre-period or the post-period after we'd done something to them? And we didn't really get into the idea of trying to look at the entire course of their cancer care and how many times that popped up or when. That's a little bit of a different beast and a little harder to control. But we wanted to know in the time around when this is happening, when they're getting spine disease, what is their kind of general state at that point? And we defined financial toxicity as kind of meeting any of those criteria from the previous slide. So bills sent to collections, bankruptcy, settlement plans, use of the financial assistance program, social work concerns, all those things. And ultimately, we were able to pull about 881 patients out of that. And about two-thirds had fusion surgeries. One-third didn't require any type of a fusion surgery. As you can see on this graph, fusion surgery doesn't matter in terms of your risk of financial toxicity. But the interesting thing is that you get about a two-and-a-half times increased rate of experiencing financial toxicity when you kind of go through the neurosurgical diagnosis plus surgical treatment with or without the radiation. So there's a big jump in terms of what's happening here with respect to experiencing financial toxicity just from us having an intervention. So it is a big toll on the patient, even if it's a small deductible that they're getting. No matter what it is, we're seeing a large increase in that. And we break it down a little bit. What we see is that it kind of comes from two ends. One is the co-pay assistance part of this. And that makes sense in this setting because in people who have new spine disease, what we're going to see is that they're going to be shifting therapies. The old chemotherapy or immunotherapy that they were on doesn't work. They're now switching to something else. That something else may be way more expensive and require assistance for paying for that. And the financial assistance program comes in because what we see a lot more of is on the back end, after surgery, needing more of the financial assistance portion of this with helping to pay for the actual bills that come from the spine surgery plus the radiation treatment plus the other things that are also going on, which makes sense. But the nice thing is no one's going through bankruptcy for this, but you can see this picture where there's a lot of people who are actually having this issue because when we look at the raw rate number, it's about 44% of the patients at Sloan are experiencing financial toxicity within two years of undergoing surgery. The downside for them is that we do have all these ways of helping them, but it's still a high number when you think about the fact that this has impact on quality of life, survivorship, ability to make it to your appointments, and to take your medication as needed. And so what are we doing about that? And I'll tie this up kind of quickly is that what we're really doing is we're actually starting to prospectively look at these people and give them cost surveys or these kind of questionnaires to see can we actually correlate the subjective experience with the monetary portion of the experience. And we're trying to track that over time to see when does this actually happen and assessing for any discordance between feeling financially taxed and actually utilizing the infrastructure that's there to get a better understanding of this and then design interventions to try to get to these people before they start missing appointments or before they start rationing their medications. With that, I will stop on time and pass it over. Thanks. Thank you. Next we're going to hear from Dr. Chang. All right, so thank you very much to the committee for inviting me to speak today. I'm Veronica Chang from Yale, so I'm going to be talking about resection-free brain metastasis management. And so, just full disclosure, nothing I'm going to talk about is cheap. All right, so the options for treatment for brain metastases are listed up here on the top, and they include medical options, obviously radiation options, and surgical options. And so, just to put this talk in perspective, if we look at our own institution, last year we did 400 radiosurgery cases and 100 operative cases, of which 80 were resective and 20 were not. And so, the reality is the majority of brain metastasis management is, in fact, non-resective. So I think that there's not enough time in 10 minutes to talk about all the non-resective options. So, the approach I wanted to take was perhaps to give you an idea of what some of the options are for what's going on with cases that we think actually should standardly have surgery. So what are our standard indications for surgery? Well, they're diagnosis, mass effect decompression, neurological relief, and then management of failure at first line treatment. All right, so let's start with diagnosis. So all the way to the left, this is usually a pretty straightforward case, 46-year-old lady with early stage breast cancer and early stage lung cancer, one diagnosed in 2012 and one in 2018, both of which could be metastatic to the brain. And so, with a single lesion in the brain that's mildly symptomatic, it seems very obvious that one would resect this and follow this with a consolidated radiation. I would propose to you, however, that there is literature out there that suggests that perhaps there are alternatives. The first could be biopsy. So there's no data out there that says that biopsy plus radiosurgery has a worse outcome than resection and radiosurgery. In fact, post-op radiosurgery is actually harder to plan, as evidenced by the fact that we now have pre-op radiosurgery trials. Late healing delays radiation and onset of next set of systemic therapy, and there's an increasing amount of literature that suggests that resection may increase the downstream leptomeningeal failure rate to as high as 35%. So is resective management the correct thing to do? In addition to that, there's new data coming out that suggests that, in fact, we may not need to even take the tumor itself at all. So this is actually work that was done by one of our residents, looking at using CSF as a way of making diagnosis. So what you can see here is that for intraparenchymal lesions, about two-thirds of them can actually be diagnosed from CSF, and that CSF, unlike plasma, can actually be a very good marker for what's going on in the brain. So then what about decompression and neurological relief? And so, again, these are standard indications, but the alternative options now include CNS-penetrating drug therapy, hyperfractionational stage radiosurgery, and combination radiosurgery with drug therapy. And so if there's any doubt that CNS-penetrating drugs work in the brain, this taught me that lesson. So this was the first case I ever saw. She's a 54-year-old right-handed white female with outgrown range non-small cell lung cancer. She was clear of disease in the body for three years and then presented with numbness in her right hand. And the MRI shows a five-centimeter lesion in the left parietal area that, for all intents and purposes, we believe is best treated by resection followed by radiation. She adamantly refused to undergo surgery, would rather die than have surgery. And so in desperation, her medical oncologist put her on ceritinib, which actually is a first-generation drug for outgrown range non-small cell lung cancer. And you can see in three weeks in the bottom row how much shrinkage she already had of the lesion. She became asymptomatic and went on to get radiosurgery for long-term control. And so this goes on with other cancer types as well, with melanoma, the combination of a BRAF inhibitor, dabrafnib, and a MEK inhibitor, trametnib, in BRAF mutant melanoma can, as you can see, result in up to 90 percent of response up front. So do we really have to resect all those lesions? And so it's not just lung cancer and melanoma, but now it's breast cancer, renal cell cancer, ovarian cancer, and the list goes on. And these drugs are coming out faster than we can predict. And many of them have CNS penetration. Moving on to radiation. So hypofractionation in stage radiosurgery is a new technique that's been around the last five years or so. So it used to be that if your lesion was greater than three centimeters in diameter, you couldn't give it radiosurgery. And that was based on the original RTOG-9005 study that showed that beyond three centimeters in size, you had to cut your dose down to 15 gray, and that resulted in insufficient control of tumor. Hypofractionation, however, is breaking up the fractions of the treatment into several doses, so anywhere between two and five doses. And so there's a lot of literature out there now that suggests that breaking the doses up can result in 80 to 90 percent local control at a year. And what's really interesting about this data is actually supported by the study by Fermenti and her group at Cornell that shows that hypofractionation may, in fact, in conjunction with immunotherapy, generate or be more likely to generate what we call an abscopal effect, which is the immune system not just taking care of the radiated lesion, but also lesions that have not been radiated around the body. And so there's some data, actually, that this may, in fact, be true. So to the left is a time graph that shows response to radiosurgery and immunotherapy. So the orange line is median volumes after radiosurgery with radiosurgery alone. And then the blue line, actually, to radiosurgery and immunotherapy combined concurrently. And so you can see that not only is there a faster and more complete response, meaning that larger lesions can be treated more effectively, but it also results in an improved survival. And so this has enabled us to actually drop the doses that we need to give in radiosurgery to treat the same lesions, so larger lesions being able to be treated again without surgery. In addition, there's also a beneficial synergy between targeted therapies and radiation, and this has been shown for multiple studies. All right, so what about management of first-line treatment failure? And so what we know is that if we live long enough with cancer, after the radiosurgery is administered initially, there's an increasing number of patients that will have regrowing lesions. And so there was some initial data that suggested that resecting regrowing lesions actually improved survival. But more recently, there's been the introduction of laser thermocoagulation. And so this, again, is a non-resective option, allows us to drill a small hole, three-millimeter hole in the skull, place a laser into the lesion, burn it from the inside. And here's a couple of cases that shows that it works pretty well. So a 42-year-old female with metastatic melanoma, you can see all the way to the left the lesion that was given radiosurgery, good response at three months, but then at 12 months it started to regrow. Biopsy showed evidence of tumor. You can see over the subsequent six months that after laser that it resolved, along with the edema, as did the symptomatology. So no need to go in and dig this out. Similarly, for adverse radiation effect, patient with lung cancer had radiosurgery to multiple lesions. You can see the right basal ganglia lesion has, unfortunately, despite the radiosurgery, progressively increased in size and became symptomatic from the edema. What's been really great about the laser thermocoagulation is that you can see that right after the laser is done, two weeks afterwards, even though the lesion itself has not changed in size, the edema from the reaction to the radiosurgery has significantly improved. She had significant symptomatic relief and actually many years of resolution. And so we know that the laser study, from the laser study, that laser works very well for both radiation necrosis and tumor. And certainly in our own institution, we looked at comparing the two tools, and what you can see all the way over to the left is that regardless of the tool that you use, if your lesion is less than three centimeters, then it looks like your outcomes may be the same, and really outcome is more dependent on pathology all the way to the right. And so I'm just going to finish up and say that I think it's important to recognize today that operative management of the standards are not necessarily the best. This is a patient with melanoma who clearly presented with gait difficulty from a very large lesion. And so there might be many of you in the audience that say that you could resect this without too much morbidity. I personally can't. But in a patient who's got a liver and chest riddled with melanoma, this patient got a third ventriculostomy, two rounds of 12-grade radiosurgery to the lesion that you see here, residual immunotherapy, and had a wonderful result. And so it is possible to do this non-resectively. It does take a team, however, so thank you to my team, and thank you for listening. Thank you. Next is Dr. Goodwin. It's actually nice to follow Dr. Chang because I think you'll see a lot of the same principles and also see where we're headed in spine. So Rory Goodwin talking about spinal metastasis management in the era of SBRT. These are my disclosures, not really relevant. So first thing we'll talk about is kind of how we got to SBRT, how SBRT has changed the current management strategy. And then I'll show a really, you know, where we're headed and then a brief case of kind of how it applies to kind of multidisciplinary management paradigm. So we know that metastatic disease of the spine affects a significant portion of cancer patients with the most common metastasis arising from lung, breast, and prostate. And some of these patients will develop a symptomatic lesion that will ultimately lead to a neurologic disturbance that ranges from pain or paresthesias to weakness and paralysis. And we know that over time that the incidence is increasing. And so you have to remember that the goal for spine metastasis management is primarily palliative with the goals of treatment basically being preservation or restoration of neurologic function, maintaining local tumor control, ensuring mechanical stability, improving pain relief, and overall improving quality of life. And so the roots of SBRT are really grounded in the PATCHEL study that compared radiation alone to radiation plus surgery with the primary endpoint of ambulation. And what this study really emphasized, at least to me, is the importance of local control and demonstrated that direct decompressive surgery plus post-operative radiotherapy was superior to treatment with radiotherapy alone. And so that PATCHEL study used conventional external beam radiotherapy, which was the mainstay therapeutic modality with typical dosing regimens, as you see here, from eight gray in one fraction to 30 gray in 10 fractions. And the advantage was that these were best for radiosensitive tumors such as myeloma or lymphoma, isolated marrow-only disease with high local control rates, and high rates of partial pain relief. Now the disadvantages of it were that you'd have low rates of complete pain relief, and it was kind of poor in terms of dealing with complex or bulky tumors. And so the question came up of, like, is there a room for improvement? And so stereotactic body radiotherapy was able to develop to give a locally ablative tumor dose in a high-dose-per-fraction regimen that was effectively five to eight times your standard EBRT dose. And although there was some initial barriers to widespread adoption, secondary to the risk of radiation-induced myelopathy, multiple studies have shown that this is overall safe. But with that said, the challenge has always been how do you maximize that cytotoxic tumoral dose, minimize that risk of radiation-induced myelopathy, and then prevent the subtherapeutic dosing of your epidural space, which is usually your most common area of failure. And so as surgeons, we've had to think about how do we want to redefine our indications and basically identify what are the areas that we can provide most benefit. And I think that focus has been on the most common patterns of failure, which is that epidural space. And so that's where the goal of separation surgery came in. And so Bilski et al., they demonstrated benefit of circumferential decompression and fusion, best seen in this paper. And then this coined the term separation surgery, which is really a procedure in which tumor resection is limited to decompression of the spinal cord to create enough space to provide a safe target for spine SBRT. And so separation surgery identifies that the normal dural planes for epidural spinal cord compression creates adequate separation between the neural elements and tumor, typically about one to two millimeters or so, basically to facilitate your SBRT and basically allows the SBRT to take care of the rest of the local tumor control, irrespective of the tumor volume. And so this approach offers several advantages ranging from no need for major cytoreductive surgery, circumferential decompression, decreased morbidity, as well as the other advantages seen here. And so from there, that led to the development of the NOMS framework. And this is really incorporating the main elements of decision-making into how we think about this and how we incorporate SBRT specifically to optimize clinical outcomes and quality of life for patients that are diagnosed with metastatic spine tumors. And so the neurologic, basically whether patients have evidence of myelopathy or radiculopathy and the degree of epidural spinal cord compression, oncologic, whether the tumor is sensitive to radiation, therapy, or I like to think also sensitive to other targeted therapeutics, mechanical stability, which is often defined based on the spinal instability in the aplastic score, which basically is a scoring system that's divided into different criteria that can tell you whether something is stable, potentially unstable, or unstable. And then systemic factors such as, you know, whether a patient can tolerate surgery, and these deal with frailty and as well as their overall systemic burden of disease. And so if you use that treatment paradigm, that NOMS framework using SBRT, you can see that, you know, multiple studies have shown that you can obtain more effective pain control with SBRT alone, but also with the combination of surgery and SBRT. This study in particular demonstrates that following surgery and SBRT, you have decreases in your pain severity, your pain interference, and other pain-associated patient-reported outcome scores. If you assess the persistence of epidural spinal cord compression, meaning how well you do in terms of actually separating it and providing that local tumor control, you'll see that if you perform a true separation, such as postoperatively, if your Bilski grade is anywhere between a zero or a one, these range far better, and that's those criteria in green, in relation to those that are the Bilski grade two or three after surgery, which they have a lot lower overall tumor control probability. And so, you know, if you examine SBRT in separation surgery, you see that we're able to decrease overall pain and then better improve our local control. But there are additional, I think, questions that are remaining in this SBRT era. And so for radoncs, I think that's, you know, can we increase our indications, provide more optimized dosing from the medonc perspective? Can we be more aggressive with our local therapies, given that our—I'm sorry, more aggressive with our systemic therapies, given our local therapies are safer and better overall with SBRT? And then from the surgeon's perspective, can we minimize the surgical footprint? And that's why I think the—you'll see where I'm going with this from the previous presentation by Dr. Chang. And so this was a recent study, just in terms of optimizing dosing from the radiation perspective. This is a phase two, three trial that demonstrated SBRT at 24 gray and two fractions was superior to conventional radiation at 20 gray and five fractions, in terms of getting you a complete pain response. And then this is a machine learning algorithm done by Soltis et al., in which they were able to show that you can actually predict your tumor control probability based on the existing studies and based on existing evidence to figure out, is there a better dosing regimen that can get us to that local tumor control that will ultimately prove better for overall our patients? And then from the medical perspective, you know, there are a wide array of key molecular targets that any person kind of treating a spine metastasis needs to know when deciding what the appropriate intervention is for a given patient. I think with SBRT we have to think about how we can better utilize and incorporate these into our overall decision-making strategies as well as overall treatment. And I think you know really what it is in terms of SBRT is can we use targeted therapies to predict sensitivity to SBRT, whether we can get away with no surgery for a patient, or conversely whether we need to offer definitive surgical treatment to maintain neurologic status to allow these patients to remain eligible for treatment. And the reason why that's important is because you know at the end of the day we need to basically figure out ways to maximize quality of life and minimize the negative sequelae, kind of like what was mentioned before by some of our previous speakers. And so really briefly you know the MIS surgery minimal access techniques, I think this is where kind of we're headed ultimately. Systematic reviews in this area have shown that your complications were less frequent with MIS compared with conventional opening surgery. Conventional open surgery you have shorter length of stay and then but no difference in your neurologic improvement. So as long as you kind of encapsulate those principles of providing that durable circumferential decompression and then giving SRS you can have overall better outcomes. And then one of the other things which is a future direction is I think lit for spine surgery. This is kind of pioneered by Claudio Tatsui and MD Anderson using kind of the same thing that you would in brain metastasis, minimally invasive, you know able to cook it in real time with MRI thermography. And I think this is going to be one of the things that is going to be for future directions for where you can see SBRT really combined with our existing therapies from the surgical perspective. And so in closing I'll just give you a brief example of kind of how I've changed my overall management paradigm in the SBRT era. This is a 55 year old male, past medical history of metastatic melanoma, widespread, bed-bound for a week, MRI demonstrated an epidural T45 met with epidural extension causing spinal cord compression. I ended up doing this separation surgery, you know two up two down classic. He had a BRAF positive mutation, ended up getting put on BRAF inhibitor as well as a PD-1 inhibitor, ended up getting conventional radiation T45 and also at L1. Ambulates one month after rehab, stable systemic disease for two years. Ends up coming back with this T11 lesion, conventional radiation, I'm sorry, and then basically that shows that he's got a SIN score 14, so suggesting mechanical back pain. The pain management doc says he's increasing narcotic requirements, he gets referred for surgical intervention. So using that same framework with SBRT, I basically incorporated some of the molecular markers in it and then basically using this, and I'll just go on, showed that he's unstable there, offered stereotactic radius surgery, and then basically ended up doing this shorter procedure to stabilize a couple of band-aids. Don't worry that's two band-aids, not four, don't worry about that. But at the end of the day, this was able to get that patient home and you know the next day ended up basically figuring out that he developed BRAF resistance, we switched his overall therapy, and he's been stable disease for the last two years. So with that said, you know, SBRT provides durable control. I think that surgery is going to be more for stability as an adjunct of treatment. We're going to move forward with minimally invasive techniques in that treatment algorithms have to identify better ways to incorporate some of our molecular markers. So with that, I will leave it with you with this, which is basically, hopefully SBRT will get us to the point where we won't necessarily need surgery, and then that way we'll be able to prevent complications altogether. Thank you. Thank you. Next we'll hear from Dr. D'Amico. Good afternoon and thank you for giving me this opportunity today, here at my disclosure. So there are a number of technical adjuncts that you can use in brain tumor surgery and low-grade resections. Those technical adjuncts, they span from the field of neuromonitoring through intraoperative imaging, intraoperative imaging, either based on pre-acquired imaging or real-time intraoperative imaging. Neuromonitoring is extremely important, as you can see. I don't know why this is advancing by its own. I'm not touching it, you see? Okay. It's magic, you know, it's technology. That's why we're talking about. So let me see if I can advance. Okay. So I was telling you about neuromonitoring. As I said, neuromonitoring is extremely important in order to identify and spare eloquent areas. We find extremely interesting and useful this device, which is a suction simulation probe, which is a suction probe, but at the same time is also a simulation probe, in a way that provides a continuous dynamic irrigation of the impulse and stimulation in a way that it's advancing by its own. I don't know why. Let me go back. Yes. So there is a continuous stimulation either at the cortical or subcortical level. Neuronavigation is based on standard imaging. It's used as a routine. Now every operating room is equipped with this kind of system. You know that segmentation and virtual reality is also useful, not only to better visualize the tumor, but also to proceed with planning of your craniotomy. Today you can also incorporate this into neuronavigation in a way to better visualize the structure you're dealing with. Either way, you use neuronavigation or augmented reality. Those are based on preoperative imaging. That means that we're talking about virtual imaging. This is not real-time. And as a matter of fact, they cannot take into account phenomena such as the brain shift and brain deformation. Interoperative MRI, on the other hand, is now being regarded as the gold standard for interoperative imaging. It's standard imaging. It's extremely accurate. The images are outstanding, as you can see here. However, it requires a dedicated area and it is time-consuming. It is extremely expensive. And on top of everything, it cannot be considered a true online device. So why interoperative ultrasound and neurosurgery? First of all, this is the most diffused imaging technique in medicine. It is extremely expensive. You can use it for almost every district of the body. And the brain has the ideal substrate for the ultrasound waves propagation. And the interest among neurosurgeons is rising, as you can tell from a growing body of literature. And this is true. Sorry, I'm smiling because he's advancing on his own. I don't know why. Let me go back again. So yes, the reason of this interest is probably the quality of the image, which has improved dramatically in the last decade. But still, this is unusual imaging. There are difficulties with the orientation. There are artifacts you need to take into account. And as a matter of fact, even for an experienced neurosurgeon, it can be difficult to identify all the anatomical structures which are depicted in a scan like this. Unless you don't have a coplanar MRI. If you have a coplanar MRI and you are not accustomed to ultrasound imaging, it's advancing, it becomes much easier to identify what is the tumor, what are the key anatomical structures. You need to keep in mind that there are some specific echoic characteristics for a single structure. For instance, the brainstem is always hypoechoic. The ventricles is always hypoechoic, since there are the croid plexus. The pineal gland is always hypoechoic. The doral structure is always hypoechoic. So those are characteristics that you need to keep in mind. And so for this reason, especially the beginning of your experience, it's better if you can exploit the integration of ultrasound scan and conventional neuronavigation, as you can see here. So here are some tips that I would like to give you. When you're doing your scan, try to scan along the orthogonal planes. This is much easier to understand what you're seeing. Consider that this is a dynamic tool, so you need to move your probe. You need to bend it. You need to twist it. You need to scroll the whole area, the whole craniotomy. And then you may want to find and define your target. You want to see the components of the target, the C6 solid. You want to see the boundaries. And then you want to identify all the nearby key anatomic structures, as I highlight here and that I already told you earlier. And finally, you may want to adjust for the brain shift. As a matter of fact, you can see here that the two images are not overlapping. This is due to the brain shift. So you just freeze one of the two images, you drag it over, and you correct for the brain shift. You do this quite a few times. In this way, you will adjust for the brain shift. Of course, at the end of the section, you want to check for tumor remnants, as you can see here. Additional tips, don't use TabuTab. Do your scan before putting TabuTab, because TabuTab will give you a lot of artifacts. Ensure a continuous irrigation and use possibly a linear probe and scanning along the edges of the resection cavity. But here we come to what is the probably the major drawback of the ultrasound. Especially when you have a large cavity, at the bottom of the cavity, you will experience most often some artifacts. Those are due to the different acoustic properties of the celling and of the brain. So this may hinder the interpretation of the images. For this reason, as you can see here, there are some tricks that you may exploit. One is to scan tangentially outside the resection cavity. Another possibility is to use a small probe, which you can stick inside the cavity in direct contact with the walls of the cavity itself. Or we are currently investigating, in a clinical trial, this acoustic coupling fluid. This acoustic coupling fluid, as you can see on the right side here, is getting rid of all the artifacts. Those are additional examples of what I'm telling you right now. I'll go quickly for the second time. So we said this is a very strictly operator-dependent technique. So there is a need for specific training. For this reason, we have formed an international group of experts and we run courses around the world. When we run these courses, most of the time we have also hands-on sessions, including simulation sessions, because simulation has been proven effective in accelerating the learning curve of trainees, as you can see from this publication. So now we have made this technology available over the internet, so anybody may log in for free and you can experience first-hand the use of Intrasound in the operating room. So here we come to the conclusion of this difficult presentation. I'm sorry about that. So there are several technological advances we've seen. Intraoperative ultrasound, we have no doubt, is the only one capable of providing real-time images. But anyhow, all technologies are useful when a variable should be integrated together. Now, some key points that I would like to highlight in regards to intraoperative ultrasound. Always keep in mind that this is a dynamic tool, so you need to move the probe. Don't use it statically. And some anatomical landmarks have specific characteristics, so you must know the characteristics of the anatomical structures of the brain. And finally, when possible, use tricks or, whenever possible, also devices in order to get rid of drawbacks. Thank you very much for your attention. Thank you. Next we'll have Dr. Prevodello. So do I click on this? Is that the way it goes? To click, start. It's just a little heavy. Here we go. Thanks so much. So continuing, the whole idea, everything we do is really like to try to preserve function. So I'm going to show some of our tricks, some of our philosophy, and some parts of endoscopic surgery when we are like emphasizing preserving function. I'm gonna show basically more focus on endoscopic endonasal, as like I was thinking about this talk, and then when you go open approaches, like we're mainly preserving cranial nerves and things that we do commonly. So I'm gonna show a little more of other thoughts when we come endonasal. So for the most part, when we do endonasal surgery, initially in my career, we did a lot of nasal septal flaps for reconstruction, and this is something that we changed over the years. We started doing what we call rescue flap for small pituitary adenomas, but we do that for all pituitary adenomas nowadays. So we don't do the flap if we don't need to, again, to preserve function. When we are preserving, the thought is to preserve the smell here. You don't want to disrupt that septum too much if you can't avoid it. So here's the this video. Do you know how to? It doesn't show the bar on the video here. I guess we wait. So we'll just move on here. So this is the way it looks like when we finish doing rescue flap. So we preserve the pedicle of a potential nasal septal flap, and the whole idea is to really preserve the anatomy, preserve the mucosa, preserve the smells and olfaction for this patient when we do these approaches. So the other thing I want to show is that a pituitary gland, the other function that we are very vigilant. So I want to show these because a lot of people don't pay attention. When you put an incision on the cell, if you look for, you see here with the Angelina disector, we are separating the pituitary gland, which is paper thin, from the pituitary denoma. So I advise you to look for the pituitary gland earlier in the surgery because that allows you to determine the plane between the pituitary gland and the adenoma very early and protect that pituitary gland and make sure that you preserve function. So we do a lot of these extracapsular resections, and I wish that this should come with a bar so I can slide here, but because otherwise it would be too long of it. I'm gonna need to escape some of these videos, but when we do extracapsular resection, we have to be very careful with the posterior gland, so we want to avoid diabetes insipidus. So one of the things that we've been using is ultrasound as well, as you saw the talk before by Dr. DiMenco, and you can actually see the pituitary adenoma, you see the pituitary gland, and you see the posterior pituitary gland as well. It's very interesting. This was my own learning curve, like if you see deep here, the white is the dura behind the cell, and you see this little black, that's the posterior pituitary gland that you can see. So this is the post-op imaging. Here is during the surgery, you can see, I've been using a dissection through the gland, as you can see this little line, you see my arrow there or no? How do you transfer the arrow? oh, here we go. I guess I have to look at it. So if you see, this is the pituitary gland that I dissected and this is the posterior gland. And you can do dissection through the gland and periodically explore with the ultrasound and that allows you, allow to protect that posterior gland. So you can see here the tumor and you can see posterior gland with the ultrasound that allows us to continue protecting the gland and function. So for large tumors, one of the concerns I have is after the resection of these tumors and this one here, I see the bar there. I guess I have to look there, interesting. Okay, so we're here at the end of these resections and here we protect the pituitary gland anteriorly. But the concern is diabetes insipidus. So you get this resection and the pituitary gland now that was close to the hypothalamus gets pushed down into the cella and the pressure of the spinal fluid stretches the pituitary stock. So we noticed that large pituitary tumors were the main reason to cause diabetes insipidus. And we look into this, we just published, just March of 2023 came in a operative neurosurgery where we measured how much is stretch on that pituitary stock. And we noticed that that stretch correlate with diabetes insipidus that could be transient or sometimes even permanent for these patients. So we start using buttress inside the cella to keep that pituitary gland as high as for a long period of time and avoid that pressure. Almost like create a contra pressure with the CSF leak. And we've just published this, you look at the last line over there and you can see here even in multivariate analysis, the pituitary stock stretch was really one of the factors that contributed for pituitary diabetes insipidus. Some other considerations on craniopharyngioma. The craniopharyngioma is a difficult surgery. I think coming endonasal is one of the advantages because allows us to have these visualizations from the midline and allows us to protect the pituitary stock that you see right there, protect the optic apparatus and allows us to make decisions on the dissection of the tumor away from the hypothalamus. So you get these resections and always with the philosophy of protecting structures. I was just talking to Linda B the other day, she's gonna present next. And one of the issues we have is really how aggressive we're gonna be with the pituitary stock on tumors that are invading the pituitary gland and the pituitary stock. So frequently we will do transpositions like this one where we try to protect the pituitary gland and detach from one side of the cavernous sinus and it'll show you it looks like this type of resection, protecting the superior hypothesial arteries as well. And here you see the hemitransposition where we detach it from the medial wall of the cavernous sinus to allow us to resect the tumor that was posteriorly located. But I was preserving the pituitary stock. I was talking to Linda B like cases that I regret when I decided to resect that pituitary stock or compromising the pituitary function causing diabetes insipidus. Have a story of a patient that was admitted six months later with an ICU with pneumonia and they intubated, they didn't know she had diabetes insipidus. Her sodium went all the way up and she ended up dying of diabetes insipidus. So I became more like less aggressive and trying to preserve this stock as much as I can. And then this type of case is immediately MRI and how it looks like five years later with a pituitary gland function preserved. Sometimes the decision making is related to the hypothalamus. This is a video that I'm gonna put at the end where I had to make that decision and the tumor was really adhering to the hypothalamus there. So there was a little bit of tumor I had to leave it here because as I was resecting this tumor I was seeing PTK on the hypothalamus and that's a sign of compromised. So here's this reconstruction that we did and you see a little bit of residual there and then eventually actually recurred there. I had to give radiation to this patient. So I'm gonna just keep going here like show these are more and more examples of the importance of this technique preserving the pituitary stock. And I wanna, how long more we have? Two minutes, okay. Finalize with some other thoughts here related to when there is what we call type two of tumor. And when you have a type two craniopharyngioma inside of the pituitary stock, I use this technique that I cut the stock on the middle open like a book. So you have stock here, stock there. And it's almost like, I feel like we're doing a dissection preserving like a facial nerve in the posterior fossa trying to preserve the pituitary stock by keeping the stock preserved in the center there. So it looks like that, you look and then we work in the middle inside the third ventricle and sometimes challenges to do the reconstruction as well. But it's instead of transecting that stock, I try to preserve it, which I think is very important. That's the post-op imaging. I just wanna finalize with this thought. Now we have BRAF inhibitors like papillary craniopharyngiomas. And this is something that changed my practice. I wanna finalize with this thought here. This you see, this is a physician that came to us with normal pituitary gland and this craniopharyngioma. Instead of coming from below, I would probably compromise his function. There was no calcification on the CT. So high chance that this is a papillary. So I actually decided to come with a craniotomy, with a small craniotomy, really more for diagnosis and small debulking. As you can see here, you see the optic chiasm and we're gonna enter through the lamina terminalis there. And just to speed up here, we enter that. And eventually there. Yeah, here we go. Then we did an endoscopic exploration, enter the lamina terminalis. And just to advance, we obtained this type of resection, which doesn't look good, but it shows the importance of preserving function. He woke up completely neurologically intact with normal pituitary gland. And then with one year of BRAF inhibitors, it evolved to this. So I just wanna show you, and this is with no radiation. So I think there's a lot of that's changing. There's a need for more trials without the use of radiation and we're trying to work on that. And with this, I thank you so much. And it's very important to preserve function in neurosurgery. Thank you. Next we have Dr. B. Wonderful. Good afternoon. It's a pleasure to join this friends and family discussion on meningiomas. And the last decade has been an exciting one. Let me make sure I just press this. The last decade has been an exciting one for meningiomas, in which numerous studies have helped to characterize these tumors, as well as to better improve our features that identify their outcome in aggressive behavior. This led to the 2021 WHO update that for the first time incorporated molecular features such as homozygous loss of CDKN2AB, as well as TERT promoter mutation into the definition of an anaplastic grade 3 meningioma. Since then, several additional studies have further clarified that both homozygous and heterozygous loss of CDKN2A2B equally portend a poor prognosis, as you can see in the parallel Kaplan-Meier curves of both heterozygous and homozygous loss doing poorly in these tumors. TERT promoter mutation, I would say, is a little bit more controversial if you control for all other molecular features, and that's still under intense investigation. But these aggressive features only capture a small percentage of malignant meningiomas, and when we look at a genome-wide scale, multiple parallel studies, including RNA expression studies by Akash Patel's group, unsupervised methylation classification in this elegant work by David Rowley's group, as well as supervised methylation classification in work initially led by Felix Sam, as you can see, and then subsequent work with Gheller-Zadeh's group, you can see in these prior curves in which the lower the curve, the less the prediction error of reality, and the reality being recurrence. So the lower the curve, the better it is. And what all these serve to show is that genome-wide molecular characterization of meningiomas can better predict their outcome. For us in Boston, we don't have easy access to methylation platforms, so we've turned to broad copy number chromosomal alterations and defined a molecular integrated grade based on these fairly transparent and easy-to-identify features that can be derived from any other molecular platform, whether methylation, RNA-seq, mutation profiling, or so on. And what we noticed in this is about a third of our meningiomas differ between their assigned WHO histopathologic grade and their molecular integrated grade, which is in line with the molecular classifications of other schemes that have been reported. What that means on a practical level is that we frequently will see discrepant tumors such as those shown here, in which the so-called malignant tumors, the blue line of WHO grade 2 to 3 tumors that were molecularly benign did far better than the so-called benign tumors WHO grade 1 that are molecularly aggressive, integrated grade 2 to 3. And using these schemes also especially gets better and better on long-term follow-up, which is consistent with the growth pattern of meningiomas. Additional recent work have increasingly showed that when you combine multiple modalities, whether methylation or gene expression or copy number, if you had all the resources in the world, you can further refine classification. But additionally, for those of us who may not have all the resources in the world, that if you had any one of these schemes, they do pretty well and are fairly comparable in this comparison between methylation versus gene transcription versus cytogenetics versus simpler schemas, that there is a high degree of overlap. So that what's perhaps most important is not having the perfect platform, but to have any molecular classification and profiling of meningiomas to augment our understanding of these tumors. But ultimately, even as we are increasingly understanding what is a malignant or aggressive meningioma, we're here to treat the patient more so than the disease. And what do we do in practical terms with a patient whose profile comes back like this, a WHO grade 1 meningioma with elevated proliferative index of 20 percent and whose molecular profile has multiple high-risk genetic alterations? Or conversely, a patient like this, which is 7.5 percent MIB1 index, and then whose molecular profile has no concerning alterations? This translates, as surgeons, to how we performed in surgery or perceived performance in surgery, including young patients who might have a fairly straightforward convexity meningioma that turns out to be grade 2, and with subsequent recurrence again and again, despite additional treatments, or in patients who, again, like this young woman, who had a classic putreclival meningioma growing over two years of observation, for which I felt that, although I achieved a fairly thorough resection, but there are undoubtedly microscopic residual cells around the cavernous carotid. And so what do we do with these patients in both short-term as well as long-term follow-up? Do we observe and reoperate when they recur? Do we offer upfront or adjuvant radiation, or do we offer targeted therapies when they do recur? Well, one thing that I will point out is that in truly biologically aggressive meningiomas, whatever we do, whether it's surgery, radiation, targeted treatment, seems to have a durability of control. And that has been shown, at least in our institutional series, again and again. So that concept of lifetime of control is important, especially when we think about younger patients whose natural lifespan may have another 30, 40, or more years that we need to think about. This is particularly salient to radiation-induced meningiomas, which we all know tend to be more aggressive, and whose profile, at least in our institutional series, shows about half of the WHO grade one tumors shown in this COMU plot, in which every column is an individual tumor, and the y-axis, or each row, represents a different chromosome. You can see the WHO grade one tumors on the left-hand side, and the majority of those have additional high-risk molecular features. More importantly, when we look at our radiation-induced meningiomas, each one of these plots is a single patient with up to 60 years of follow-up. And you can see about half of these patients, once we start intervening, are unleashed in the number of interventions given, whether it's surgery for the tumor, surgery for the cavernoma, plastic surgery for the wound revision, the shunt, or so on. And so that lifetime of potential interventions for a young patient with aggressive meningiomas is something that I think is important to temper as we increasingly uncover their potentially more aggressive behavior up front. And considering the timing of interventions, especially when there is asymptomatic residual or recurrent tumor, to maximize the duration of a control so that we're treating the patient and their mortality and morbidity over the decades, as opposed to treating a radiographic picture, is probably the single most important thing we can do in considering tumor factors for meningioma. Thank you very much. Thank you. Our next speaker is Dr. Drummond. So, I'd like to thank the organizers for giving me this topic. I was a little bit interested at the beginning. I've never made a nomogram in my life, however, they were quite prescient because I've suddenly worked out exactly how it is related to the work that I've done, so I've enjoyed putting this talk together. So the title was, Can an Integrated Nomogram Provide Better Care for Glioblastoma Patients? The answer, if you want the short version, is yes. If you want the slightly more nuanced version, it's sort of, and my disclosures are just some grant funding for some of the work that I discuss at the end. So there have been a number of nomograms put out for various risks for our glioblastoma patients and they're great because they can look at the risk for an individual and you can look at different outcomes, overall survival, progression-free survival, quality of life at 12 months, you can decide what you want to put in. They can be used for treatment decisions, planning, prognosis, follow-up, patient counseling, clinical trial stratification, and you can have a little online calculator. So this one here is very nice. It uses age at diagnosis, gender, KPS, MGMT, and resection status to tell you the likelihood of survival, prediction survival at 6, 12, or 24 months. It's got a nice little online calculator, so it's great to use. Of course, there was one made with the population from the large NCIC EORTC trial, the Stook trial, which came up with fairly similar predictors. You could use the whole population if you wanted to and you could look at either survival probability at 2 years or median survival, but you could also change the groups around, which is one of the important things about these nomograms. If you only looked at the patients who had MGMT status available and who had tumor actually removed, which wasn't the case for all of the patients in that trial, then you came up with a much smaller group of predictors in your nomogram, performance status, MMSE, and MGMT, so that changed things a lot. The utility of these nomograms depends on multiple factors. The first one is a good question being asked, and is the outcome well-defined? That's fairly obvious. Which patients were included and do they represent your patient? The majority of these are based on clinical trial groups, and as my work has shown and other people's work has shown, clinical trial patients do better than your average patient. So were patients included who were of the age that you're looking at, of the performance status that you were looking at, did they all have surgery? Which input variables were included and how were they defined? How was MGMT measured? How was extent of resection measured. In many of these trials it's surgeon's opinion, which as we know is a little overestimate of how much tumour's been removed. You know, were patients with biopsy only put in? How many and which data sets were used and what was the validation? Was there an external data set used if it was single institution? This one's a big one. Does the user understand the statistics? We all look at the papers and we go, oh, area under the curve, 0.75, that looks good, that was included. But that only means that it can predict an event for a patient having or not having the event 75% of the time. So does the user understand? Does it include variables important to the patient? And are the numbers and risks generated helpful or harmful for the patient or for anyone to know? They're all interesting things. So something like this one, which is a real world group of patients with many different treatments and not a clinical trial may be useful, but interestingly, still comes up with probably the same group of factors in the nomogram, age, resection, IDH, radiotherapy, chemotherapy, et cetera. An integrated nomogram of course is one that includes usually molecular data, often adds a gene signature. That becomes somewhat problematic because you can put like genome wide sequencing, you could just put in a couple of known variables. You could put in pathway features that you think are risk but that you don't know are risk. So potentially prognostic genes. So that can change things. This is one that in the end used IDH, P10, and P53, which is probably something that we're all doing in our own heads anyway most of the time. So I think an interesting question though, and which comes to some of the work that I've done, which is why I was asked to give this talk, was can an integrated nomogram predict the risk of getting a GBM? Not just treatment outcome, but can it predict the risk of getting a GBM? So we have this group, AGOG, the Australian Genomics and Clinical Outcomes of Glioma Group. We've got linked clinical outcomes resource and biospecimen repository. This was a large familial study, largely looking at risk. So 600 participants with 400 family members, usually siblings, but if not siblings then spouse. So siblings of course for genetic information, spouse for environmental factors, blood or saliva, multi-institutional across Australia and that's closed to accrual now and some of the data is starting to come out of it and maybe this is where I'll build my first nomogram. One of our interesting papers was that the data that we collected from these was very wide in terms of lifetime risk of all sorts of things, like known risks, radiation, genetic diseases, familial risk if you've got a first degree relative who's had a GBM, you've got a two-fold risk of getting one yourself. So all of that, but physical activity, occupation, all their jobs, everything over time. The first one was that we've probably mostly missed the boat in this room, but if you were very active in adolescence then you're much less likely to get a GBM. So if there are young people in the room, young adult probably works as well, start exercising. But also your lifetime activity will probably change your risk for having a glauoma. And then another interesting one, and don't ask me about how to do genome-wide association studies, but some people in our group know how to do this. So the genome-wide association study which really showed, validated a number, there's 50 risk SNPs for glauoma in 34 genetic risk regions, and validated some that were already known, but showed a very strong risk for women with this looking at more than 7 million SNPs looking at this particular region. So it may be that at some stage we're able to put together a nomogram where you can very helpfully look up your lifetime risk of getting a glauoma, and I'm sure that will all be right on there to do that. Thank you very much. Thank you. Next we'll have Dr. Nahed. Great. Thank you. Thank you to the organizers and the moderators for the invitation. It's an honor to be here. I'm at Mass General. I'm going to present a little bit of our work on liquid biopsies, and hopefully none of this is new. And, you know, as was heard earlier today, hopefully an exciting future for the role of liquid biopsies with brain tumors. I've been tasked with talking about them in recurrence. I have no relevant disclosures to the talk. In looking at liquid biopsies, particularly in brain tumors, it's helpful to do it in the context of the successes and failures of a lot that's been happening in the other cancers. Hopefully I'll leave you with the notion that it's not just single one single aspect of it, but rather a multi-assay liquid biopsy, which we're doing. But ultimately I think we're all here to say, is it really time? Are we there yet? Are we able to use this for patients? And if you look at just this plot of the number of papers that have come out in the last five years, it eclipses everything that started when I first got into this field, not even more than 10 years ago. And there's a lot to be successful or proud of, of the successes. The first is our diagnosis, right? We've gone from looking at H and E to now looking at pathways, looking at mutations, looking at different aspects of the tumor. We've taken that diagnosis. We can come up with a better prognosis because we can incorporate that. The WHO accordingly has moved. Ultimately it comes down to extent of resection, but it also comes down to the type of molecular markers. So the importance of knowing exactly the type of tumor we're treating really affects the prognosis. If you look at the heterogeneity, this is all the work that this group has put out. You can see that we're not just talking about one tumor, right? If you look over here, this is the primary GBM. It undergoes surgery. You underdo some type of therapy and then ultimately you get a recurrence and there's a second surgery. And yet we have tissue from this point and this point. And yet this vital portion in the middle is really where the role for a liquid biopsy probably has the greatest significance. Everyone knows this sort of pathway. The patient comes in, you get the MRI, you get the surgery. And again, that defines so much of the care, but it's kind of ridiculous to think about that care dictating how that person is treated a year, a year and a half, two years later. It's even more crazy to think about the fact that at this stage we make decisions as to whether or not therapy is working based on an MRI that we know is prone to issues. Now our field has looked at this and we have guidelines around the value of recurrence, surgery of recurrence. And you can see, again, for diagnosis it's great, for symptoms it's great, for even some of the quality of life and the KPS scores it's great. But ultimately it gives us tissue. So what about the role of a liquid biopsy? And hopefully I've left you with the idea that not only would it be valuable at the initial, throughout therapy, but most importantly at recurrence. So when we talk about liquid biopsy we really talk about three different sort of tranches. The first is the actual cells that the tumor releases into the blood. And it's long been known that this actually exists in gliomas despite everything we're taught, and they're called circulating tumor cells. They're representatives of the tumors. Some would argue that they're a different portion of the tumor that breaks free, but ultimately it's still a glioma cell. The second one is the portions of the tumor from apoptosis or some type of cell event that's scattered whether it's in CSF or blood. And then the third is the intentional release of these microvesicles that happens with all types of cells, but particularly with tumors. Now the value in this is that the CTC has absolutely everything that the tumor has. The problem is that it's rare, it's fragile, and only a small group of people can do it. The value of circulating DNA, RNA, any type of aspect, messenger RNA, is that it's abundant, but the problem is you got to get to it pretty quickly and it's super fragile. And then extracellular vesicles have really taken off because they're abundant, but they are fragile. So what do they look like? Well this is our work on circulating tumor cells. This is the first evidence of one in a patient. It's an adult patient with a GBM, and you can see it's very similar to the other types of cancers, both melanoma, prostate, lung, breast. They usually are quite large in the sense of they're 10 microns, so they're bigger than things like platelets, but the hallmark characteristics of them are very, very similar, an abnormal looking cell. Now the problem is this is pretty hard. Since we put out our first paper, only five groups have replicated it, and the detection rate is over here. You might be able to read that, but it's all less than 50% of patients, even in the best of hands, making this, while valuable, something that is something to aspire to. Now what about some of the free DNA, the free RNA in gliomas? There's a lot of success in this, and you can see this chart in the bottom is just some of it. TERT, IDH, BRAF, you can find a lot of these things, both in blood and in CSF. With detection rates well above 50%, some would argue even into 75 and 80%. Now if your biofluid of choice is CSF, you would say that that's even better because you don't have to sort through the stuff in the blood. However, if you're a blood-based biofluid person, you would say that ultimately it's hard to get an LP or to convince a patient to undergo that. And then extracellular vesicles, these are abundant. And again, it's part of the normal part of any cell to release these. With tumors, they're incredibly valuable. And this author over here on the left actually speaks of extracellular vesicles differing from different parts of the tumor. And so you can imagine, if you're able to capture any one of these, you can all of a sudden start to talk about the tumor in a very different way, particularly as you're undergoing therapy. When you put it all together in this review, you can see that absolutely every single aspect of a liquid biopsy has some strength. But it's really got to be balanced with the limitations. And that limitation ultimately has probably prevented it from being used currently and hopefully won't be a barrier in the future. So what does the future look like? Well, this is how a lot of people are now starting to think about this. It's not an either or, but rather, how do we have them both work together? You can see here in the top right, the peripheral blood would be where you'd get some of the CTCs, the exosomes, and the cell-free DNA, not or. Similarly, if you're lucky enough to get CSF, you can get something very similar. But ultimately, this idea that you can follow a cancer throughout its progression, throughout its therapy, and hopefully its response is the elegant portion of the liquid biopsy. Our laboratory, several years ago, jumped onto this multi-assay biopsy, largely out of need, because not every patient has CTCs. And so we have, this is Dr. Shannon Stott, who's my co-PI and a brilliant engineer and scientist, and actually developed many of the assays that we'll be presenting. Ultimately, you see a patient with a left temporal glioma. They undergo surgery, but we're able to capture their circulating tumor cells. We're also able to capture their exosomes, cell-free DNA, and RNA. You can also look at expression profile. And while all of that is really interesting, this is really the crux of it. So you can see this patient is doing well, and then has this peak, ultimately has surgery for recurrence. The numbers then drop. This is CTCs on the y-axis. And then ultimately, we'll have a failure of therapy again. What's really cool is this is some of the expression profile data, as you watch the different portions of it move as people fail therapy, likely in response, or frankly, to provide that escape mechanism for that tumor. So ultimately, the role for liquid biopsy isn't really an if, but when. You can see that it really has value at any single time point, and hopefully longitudinally, because a lot of the stuff we traditionally did with tissue, we can now do with blood-based or CSF aspects. The great potential for that, I hope someone's okay, the great potential is ultimately that we could monitor and detect the role of not only our own therapy, but decide the validity of many of the clinical trials that our patients are on, because ultimately, looking at an MRI scan is too limiting. And looking at any one aspect of the liquid biopsy is also limiting, because just because the signal's not there doesn't mean that there's not a signal. It just means that you're not looking at the right test. With that, I just want to thank our group, and thank you for the time. So, for our first speaker is Dr. Xu, who is the award winner for the Leica Surgical Tumor Award. Hi, everybody. I'm very honored to have this opportunity to present my research. I'm a research fellow at the Loyal and Edith Davis Neurosurgical Research Lab at Baroneurological Institute under the mentorship of Dr. Pruill, who is the chair and director of the lab. My topic today is intraoperative in vivo confocal laser endomicroscopy imaging at glioma margins. Can we detect tumor infiltration? These are my disclosures. This is the confocal laser endomicroscopy, otherwise known as CLE, device that we used for the study. It's the first FDA-clear device. It produces intraoperative, real-time, cellular-level imaging with the help of fluorescein as the contrast. This is the first study to look at the CLE at glioma margins. In order to do that, we included CLE images acquired at glioma margins with a matching tissue biopsy, H&E biopsy, from two institutions, Baroneurological Institute and Byrne University Hospital. These images were analyzed by four neuropathologists using a newly proposed scoring scale. The inter-rater reliability, the concordance rate between CLE and H&E pathology, and the diagnostic performance of CLE was analyzed. This is the proposed scoring system. It's a zero-to-five scale, with the higher scored, the more pathological feature in the images. All of the regions of interest we include in this study were divided into a low-tumor probability group and a high-tumor probability group, based on the median score of the H&E score by the four neuropathologists. In total, we included 11 cases of newly diagnosed gliomas, comprising of 25 regions of interest, and 17 cases of recurrent gliomas with 31 regions of interest. These are some of the examples of the CLE images with different scores, with the corresponding H&E images. First, we define the inter-rater reliability of a scoring system by calculating the inter-class correlation coefficient. Overall, when used with H&E images, the scoring system has excellent reliability. When used with CLE images, it has moderate reliability. Tumor difference was found between low-tumor probability group and high-tumor probability group, and between newly diagnosed glioma and recurrent gliomas. Then we calculate the concordance rate between the CLE and the H&E pathology. Overall, the concordance rate is 61.6 percent. However, when comparing low-tumor probability group and high-tumor probability group, the low-tumor probability group has a significantly lower concordance rate. No statistical significant difference was found between the newly diagnosed and the recurrent group. Finally, we calculated the diagnostic performance of CLE at glioma margin. Overall, CLE has relatively higher sensitivity and positive predictive value in most scenarios, while the specificity and negative predictive values are lower. When comparing the two, the low-tumor probability and high-tumor probability groups, the low-tumor probability group produced higher negative predictive value, and the high-tumor probability group produced higher sensitivity and positive predictive value. And when comparing specificity of the newly diagnosed group and the recurrent group, the recurrent group has significantly lower specificity. We believe this is due to the current nonspecific fluorescein does not allow for discrimination between tumor cells and non-tumor reactive cells in the recurrent treatment-affected cases. From all these results, we conclude that CLE may detect tumor recurrence at glioma margins, but it's not highly dependent at this current stage. The scoring system that was proposed has excellent reliability when used to score HNA images, but moderate reliability when used to score CLE images. We believe that specific binding and even tumor-specific fluorophoresis may increase the specificity of CLE, and a CLE image atlas and a consensus guideline for interpretation may improve the inter-rater reliability of CLE. And finally, the role of the neurosurgeons in the interpretation, in the interoperative interpretation of CLE images needs to be assessed. Next. Thank you. Thank you. Next we have Abu Al-Ashar, who has the Lunsford and Lenskill Radio Surgery Award. All right, good afternoon. Thank you for giving me the opportunity to present. My name is Samuel Alshar. I'm a fifth year resident at the University of Pittsburgh. Presenting on the stereotactic radiosurgery for vestibular schwannoma and neurofibromatosis type two. These are our disclosures. As we know, neurofibromatosis is a rare autosomal dominant disease with a very low incidence in the population, with a hallmark of the presence of bilateral vestibular schwannoma in these patients. Multiple management paradigms have been proposed, including the use of stereotactic radiosurgery. However, there has been a great controversy regarding the use of stereotactic radiosurgery in these patients, because there is a theory that these patients are gonna be now affected with their second allele based on the radiosurgery, and that will induce new tumor development and malignant transformation. But that has been actually shown to be false, with multiple studies showing that actually patients with neurofibromatosis, they already have biallelic mutations. And when they looked into the safety of stereotactic radiosurgery, in terms of the new tumor development and malignant transformation of vestibular schwannoma, the fact that was there actually a very, very, very low incidence of new tumor development less than 0.5% or malignant transformation at risk, and that was reported in many, many studies. And therefore, because of all of that controversy, there actually have been very scarce data out there in the literature that looked into the role of stereotactic radiosurgery to manage vestibular schwannoma and neurofibromatosis type two, where the majority of studies encompasses only less than 50 patients, and limited data and no real consensus on the utility of that. So we thought that we wanted to look at to an international multicenter cohort analyzing a large group of patients with neurofibromatosis type two, analyzing the vestibular schwannoma and what's the utility, safety, and efficacy of stereotactic radiosurgery for that. So we had more than 250 patients having 328 vestibular schwannoma, and we analyzed the tumor control rate, the freedom from additional treatments, serviceable hearing preservation, and were there any episodes or incidents of radiation-induced new tumor development or malignant transformation. So here the cohort characteristics with a median age of 31, and the first factor is the tumor control rate, and what we saw is that in the 10 and 15 year, tumor control rate was around 77 and 52% respectively, and the major driver for that was actually tumor volume for the tumor control. What was interesting is this exact slide, and that's the most important slide in this talk, is the freedom from additional treatment. So stereotactic radiosurgery offered the freedom from additional treatment in the 10 year and 15 year after receiving stereotactic surgery of a rate of 85% and 75% respectively, and the major driver here again was also the tumor size and the tumor volume. When we looked at our hearing preservation rates after stereotactic radiosurgery, it's known whether you do nothing for these tumor, you operate on them, or you do stereotactic radiosurgery, the hearing preservation rates will decrease, and that's the natural history of the disease itself, but here after stereotactic radiosurgery at five and 10 years, the serviceable hearing preservation was at 64 and 35%, and the major driver was actually the patient's age and the bilaterality of the tumor location, and the second interesting factor confirming all of these previous studies is when we looked at this largest cohort present in the literature right now of these patients, we did not observe a single episode of malignant transformation or new tumor development after stereotactic radiosurgery, so that did not happen. So in conclusion, although the tumor progression rate was 48% at 15 years after stereotactic radiosurgery, what's important is that stereotactic radiosurgery offers a very, very long freedom from additional treatment with less than 25% of patient needing that, and the tumor volume seemed to influence those factors in terms of the tumor control and freedom from additional treatment, while older age and presence of bilateral tumor would influence hearing preservation rates, and we did not observe any episodes of tumor, new tumor development or malignant transformation. A lot of people I would like to acknowledge, but at the time of restraint, I would like to acknowledge our great research fellow, Dr. Othman Benammer, and my colleagues, and under the mentorship of Dr. Lunsford, who has been a great mentor to me over these years. Thank you very much. Our next speaker is Dr. Moniz Garcia. Good afternoon everyone, my name is Dior Garcia, I'm a research fellow at the Mayo Clinic under the mentorship of Dr. Q, and today I'll be presenting on rapid high-throughput identification of gliomas using desorption electrospray ionization mass spectrometry. So this is work we actually have been developing over more or less now two to three years using ambient spectroscopy. So ambient spectroscopy, just in short words, is very similar to the mass spec most people are used to, with the big advantage that the sample doesn't get consumed. And that is a huge advantage to us because those smears that I'm going to talk about can be then sent to pathology so we can have a comparison to gold standard. So you have a ground truth for making a comparison. So how does it work? What you see here on the top of the screen is a smear that we can make right outside the OR using a machine that we have actually right outside of our OR suite, which is here below. This gets read, so there's a solvent that gets sprayed, ionizes the superficial layer, bounces back at a critical angle, gets captured by a linear ion trap, and you get a spectra rather quickly. So initially our goal was quite simple. We started small. We wanted to be able to, can we detect a very unique oncometabolite that will only be present in certain tumors, which is 2-hydroxyglutarate and IDH mutant gliomas, and can we make a prediction about the genotype? Is this patient wild type? Is this patient mutant? So how did we do this? What you see here is just our workflow. On our image guided navigation we collect the tissue, get smeared right outside the OR, and within three minutes we can give you an answer. Is the genotype IDH mutant or is it IDH wild type? But this, at least today, doesn't have quite the relevance you would expect for a surgeon, so we want to ask the question, can we go beyond this? And in this particular subset of patients, IDH mutant gliomas, can we give you a sense of if there is tumor beyond the margin? So if there is a tumor, there will be 2-hydroxyglutarate. If there isn't, there won't. So what you see here is that just using that quite simple exercise, once the surgery is done, a surgeon can no longer discern whether or not there is more tumor. He cannot see anything and everything looks normal. Just take a small margin biopsy, send it to us, and we can give you within three minutes if there is an answer. Is there still tumor I can capture or not? And as you can see here in our results, we're able in 88% of cases to tell you there's still tumor, which if it's a safe area, perhaps the surgeon can continue resecting until we get you a negative result. But this still has limitations. Number one, it only applies to IDH mutant gliomas, and number two, we can tell you there's tumor but not how much of it. So we had to go beyond this. So at Mayo Clinic Floral, we do have quite a large biobank of tissue. This is all fresh tissue that we have biobanked over the last few years. As you can see here, almost 12,000 samples. So we wanted to take advantage of this now to look at the lipids using mass spec to see if one, can we make a diagnosis, meaning which types particularly of tumor? And two, can we tell you how much tumor there is actually in that Martian assessment? So we built now a robotized arm that could do the same analysis but much faster. This was quite the jump in terms of quality and quantity. We can use now just literally 10 nanograms of sample, and under one minute, we can give you up to 384 results for each of those wells you can see here. So this is our study design. We used a discovery set of 36 samples that had multiple types of tissues, normal tissue, pituitary tumors, gliomas, meningiomas, and then a validation set with 30 samples. And our question was very simple. Can we actually determine whether or not we can identify the class to which it belongs? And what you can see here is that there is a clear separation of meningiomas and pituitaries, no surprise. A bit harder to discern between gliomas and normal tissue, which you would expect given its infiltrative nature, there will be always normal tissue mixed with the glioma, but nonetheless, you can see here our prediction ability, 100% for pituitaries, 100% for meningiomas, 85% for gliomas. And then finally, the most important question to us was, can we tell the surgeon in this Martian assessment if there is a lot of tumor or just a little bit of tumor, which can factor in the decision of continuous action beyond until he thinks it's safe or stopping. And what you can see here is, even though there's not a unique peak that will identify a low tumor versus high, TCP meaning tumor cell percentage, there is a unique profile. So of course, it is a bit harder when we get to the weeds of discerning between moderate high or high and moderate, but it is quite easy for us to discern between just categorical. Is it low or high tumor? And you can see here clearly the separation in the PCA analysis. So in conclusion, we have developed a new tool that allows us to do a margin assessment in IDH mutant gliomas, taking advantage of a large biobank with more than 400 unique primary brain tumor samples and using their lipidome, we're able now to do intraoperative diagnosis and also distinguishing between high and low tumor percentage, which we think in the future will change practice. So with this, I'd just like to finish by thanking the team, I'm just the face here presenting today, and particularly Dr. Kiu, who's my mentor, and the NIH for funding this work. So thank you. Next, we'll have Dr. Casaus, who got the Southeastern Brain Tumor Foundation Award. Hi, good afternoon, everyone. My name is Joshua Casaus. I'm a PGY-5 at UCLA. I want to thank you all for being here and thank the committee for the award and the opportunity to discuss some of our preliminary work looking at EZH2 inhibition in meningioma. So first off, no disclosures. I don't have to tell this audience a little bit about meningiomas, you guys are well-versed, but what I will say is we know high-grade meningiomas invade, and they show propensity to recur, and additionally, there is a need for further adjuvant therapies for these lesions. So a little bit about EZH2 in meningioma. I'll define EZH2 shortly, but meningiomas recently has been well-versed in the talks earlier today. There's been an explosion of data and studies looking at the genetics and epigenetics in meningioma, and one of the things that has come up is EZH2 both at the protein and at the molecular level, and this study out of Yale shows that essentially EZH2 has been found to be upregulated in higher-grade meningiomas independent of NF2 status. Furthermore, when you look at EZH2 expression at the protein level, it correlates with propensity to recur, but also prognosis in general. So what is EZH2? Essentially, it's an epigenetic modifier, and what it functions to do is it methylates regions of the genome, specifically H3K27, and functionally silences these regions of the genome. From an oncology perspective, you can think of this as a tumor suppressor as well. However, it's not that simple. EZH2, depending on the type of tumor, depending on the lesion, has been shown to potentially have an oncogenic effect as well. For the purposes of this talk, we'll focus on the tumor suppressor pathway. So in terms of our objectives, simply we sought to assess the effect of inhibiting EZH2 using several in vitro models. We assessed several EZH2 inhibitors, and most of the data that I'm showing you today is from assessing the effect of tazimidastat, which is an FDA-approved EZH2 inhibitor that's well-tolerated and has a favorable clinical pharmacodynamic profile. The first question we asked is, if we take high-grade and low-grade meningioma cell lines and treat them with this EZH2 inhibitor, what happens? And for this example, I'll show you IM Lee is a high-grade, aggressive, very aggressive grade III meningioma cell line that you can see has a significant dose-dependent reduction in proliferation throughout the course of this experiment. When we look at the less aggressive cell line, SF1335, we see similar findings as well. We further characterized other meningioma cell lines and see a dose-dependent response at varying IC50s. Next question we sought to ask was, what happens when we take primary patient-derived cultures, both NF2 wild-type and NF2 mutant, and we treat with an EZH2 inhibitor? And this first, as you can see, these first three cell lines are examples of NF2 wild-type, and we see a strong dose-dependent, or a strong response across the board to EZH2 inhibition. Next we looked at NF2 mutant, and we see a significant response as well, however this seems to be a little bit more of a graded response, and when you plot the IC50s, you see that they tend to be on the higher end of the curve. This is something we're investigating currently as well. When you look at these cells under the microscope, you see that they change morphology, they senesce, or they have that picture of senescence. So one of the questions we asked is, what is happening to these cells in terms of cell cycle and cell cycle progression? And to address this, one experiment we did is we used flow cytometry to stain for the percentage of KI67 cells in this cell line. We found, ultimately, that we have a significant reduction in KI67 at 72, as well as at 48 hours. Furthermore, this was corroborated by looking at cell cycle progression by measuring the percentage of G2-positive cells after treatment, and what we see is a significant decrease in G2-positive cells across the board with treatment. And when we look at markers of cell cycle progression, or players in cell cycle, such as cyclin D1, and we look at protein expression pre- and post-treatment, we see a decrease in regards to cyclin D1 expression. And finally, we asked, what is happening to these cells? Could they be undergoing apoptosis? And we used flow cytometry and the Annexin V PI system to look at, essentially, all apoptotic cells as well as early and late apoptosis. And if you look at all apoptosis, we see a significant increase in apoptosis across the board in our cell lines, as well as patient-derived cell cultures. So in conclusions, what we've shown today is that we have a significant decrease in cell proliferation, we have a reduction in KI67-positive cells, and we have an induction of apoptosis. This is the framework for ongoing studies in the lab. Some of these studies include in vivo validation, as well as validation in 3D organoid models. Further mechanistic studies, and really based on some of the exciting talks we heard today, the questions we're asking is, what happens at the mRNA level when you treat pre- and post-treatment, as well as potentially at the methylation profile? Can we affect the profile of these tumors? Finally, evaluation in combination therapy. Ultimately, the goal is likely this will be part of combination therapy, and there is good data in other tumors and other cell lines to look at showing that EZH2 inhibition in vitro has radiosensitization effects. And with that, I'd like to thank my mentors, Dr. Everson, Dr. Giovannini, and everyone. Thank you. Our next speaker is Dr. Tang, the winner of the Brain Lab Neurosurgery Award. Good afternoon. It's my great honor to be here to share our recent research progress on the double role of the KL4 gene on the malignant progress of meningioma, which was controlled by the Link IAE. So hold on. Next. Okay, thanks. Okay. As we know, meningioma are the second most common intracranial brain tumors, and there was very diverse pathological types. And in the latest fifth edition of the WCS tumor classification showed introduced major advances that are the role of the monocle diagnosis in the meningioma and the target therapies. And this picture is provided by my friend, Linda. Thanks. Okay. This shows the totally 15 subtypes of meningioma, and which classified into three W2 grades. And last year in China, the Chinese Medical Association also issued the expert consensus statement about the monocle diagnosis and treatment of meningioma in China. And in our recent editorial also discussed the importance of the monocle genetic alternations in meningiomas. And in our last paper, we showed that the KL4 gene was a tumor suppressor during the progression of the malignant meningioma, Y, or the regulation of the tumor suppressor, protein P53. And we all know that the KL4 gene is considered to have both functions of a transcranial activation and a suppression, and as oncogene and anti-oncogene during the tumor pathogenesis, and which was dependent on tumor and microenvironment. However, how it exactly exerts its role in the pathogen meningioma is still unknown, which we think maybe involved the participation of microRNA and the link RNA. So first, we performed some of the new surface experiments to demonstrate the relationship between these factors. And the two pictures show that there was a binding site of microRNA in the P53 gene, and the microRNA can negatively regulate P53 protein expression through the binding site. And then we proved that the KL4 can positively regulate microRNA and the P53 protein also by its binding site. And also, in addition, there were multiple binding sites on the link RNA for the microRNA region, and the link RNA can inhibit the effect of the microRNA on the P53 protein expression. So next, in the tissue level, we find that the microRNA can negatively regulate the expression of the P53 protein in the benign meningioma tissues, but it has lost the function in the malignant ones. And then in the cellular level, in the benign and malignant, we selected the fibroblastica and the endoplasmic meningioma as example cells, respectively. We found that in the benign meningioma cells, the KL4 gene can downregulate P53 expression, while in the endoplasmic meningioma cells, it can upregulate P53 expression. Then we furtherly do the in vitro meningioma cell experiments from the perforation and apoptosis and the cell cloning formation and the cell invasion. And we found that for upper part, the KL4 gene acts as an oncogene in the benign meningioma cells, while it exerts tumor suppressor activity in the malignant meningioma cells, among which the INCA-IA is the key mediator for the low switch of the KL4 gene. And finally, we also performed some in vivo meningioma experiments in mice. And we've proved that the INCA-IA can enhance the effect of KL4 gene in the inhibition of the proliferation of malignant meningioma tumors and the prolonged survival time of the mice, okay? So in conclusion, we speculated that the KL4 gene maybe have the two roles during the progression of the meningioma. In the early stage, in the benign meningioma, it was a low level of INCA-IA. The KL4 gene could regulate P53 expression by the INCA-IA, while in the high grade stage, with a high level of INCA-IA, which can block the microRNA expression. So the KL4 gene can directly regulate the P53 expression. I would like to thank my mentor, Professor Gong Yu, of the tumor team research, and also the Director of Development, Professor Zou and more. And I also attach my email address here for who may be interested in this research and for more discussion. Thank you all. Thank you. Next, we have Dr. Arrieta Gonzalez, who is the winner of the Ronald Bittner Award on brain tumor research. Okay, good afternoon, everyone. My name is Victor Arrieta. I am a postdoctoral research fellow at the laboratory of Dr. Adam Sonovand at Northwestern University. Today, I'm going to present you a project entitled Enhanced Delivery of doxorubicin and PD-1 blockade antibodies through an ultrasound-mediated blood-brain barrier in gliomas. So one of the things that I want to mention first is the lack of efficacy of PD-1 blockade in recent phase three clinical trials with recurrent GBM patients. This is important because this emphasizes the need to find better treatment strategies to improve responses to immunotherapy. So one of the options that has been tried in immunotherapy-resistant tumors such as triple negative breast cancer is the combination of immunotherapy with immunogenic chemotherapy such as doxorubicin. For instance, in this clinical study, these authors evaluated different treatments in combination, followed by PD-1 blockade, and we can see here that doxorubicin was the highest. So this treatment achieved the highest objective response rate among these treatments. So for this, to employ doxorubicin in GBM patients, we started to, because we know that the blood-brain barrier represents a challenge for effective drug delivery, we used ultrasound or implantable ultrasound devices in GBM patients to deliver chemotherapies and also immunotherapies in the first phase one clinical trial performed in the United States. In brief, what we do is to inject microbubbles and drugs intravenously. And once these microbubbles and drugs are circulating, then the ultrasound device emits ultrasound waves, and this induces the vibration of these microbubbles to then disrupt mechanically the blood-brain barrier. So then immunotherapies or chemotherapies can cross the blood-brain barrier and then treat the peritumoral brain regions or non-enhancing tumors. So here we can see an image where we can see gadolinium enhancing in the region of blood-brain barrier disruption. So in our first phase one clinical trial, we enrolled patients in which they have the implantation of the ultrasound device, followed by treatment with albumin-bound paclitaxel. Many of these patients progress, and in order to provide an additional treatment, we started treatment with doxorubicin and P1 blockade, and then these tumors exposed to these therapies were resected and then further analyzed. Thus, we had per-analysis, pretreatment, and on-treatment tumor samples for the same patients. In addition, Dr. Sonovand, he also took peritumoral brain regions. You can see in these images regions that fluoresce, which we can induce this injection with fluorescing, and then we took also peritumoral brain regions that do not fluoresce, indicating regions that are not sonicated and with this, we evaluated whether doxorubicin had a higher concentration in these regions, and we found that sonicated peritumoral brain regions had increased doxorubicin concentration. In addition, we use multiplex immunofluorescence to evaluate the immunogenic effects of doxorubicin in pretreatment and on-treatment tumor samples. Here we can see SOX2 as our tumor cell marker in light blue and HLA-AVNC in orange. We can see that on-treatment GBM samples have increased cell density of tumor cells expressed in HLA-AVNC, and this was similar for HLA-DR, a component of MHC class two, suggesting increased identification of these tumor cells by T cells. In addition, we wanted to investigate the effects of doxorubicin in the tumor immune microenvironment, specifically myeloid cells and microglia, and here we can see that microglia, which you can see them here in white, they produce more interferon gamma, and these were tumors exposed to these therapies. And lastly, we wanted to also investigate whether ultrasound increased the concentration of PD-1 blocking antibodies in sonicated peritumoral brain regions. Here with two patients, we found an increased concentrations of pembrolizumab, and this was done using ELISA, specific for pembrolizumab. Lastly, we wanted to investigate whether T cells isolated from brain tumors treated with ultrasound doxorubicin and PD-1 blockade had an activation state. Specifically, we investigated interferon gamma, and we found that T cells isolated from these tumor samples had increased production of interferon gamma, specifically for CD4 and CD8 T cells. So I want to leave you with these three main points, that opening of the blocker through ultrasound increases the concentration of these anticycling and immunotherapies in the brain, doxorubicin upregulates antigen-presenting molecules in tumor cells, and also that enhanced delivery of doxorubicin and PD-1 blocking antibodies by ultrasound induces an interferon gamma phenotype in microglia and T cells. And I just want to thank, and also for your attention, also my mentor, Dr. Adam Sonovan, for being an amazing mentor. Thank you. Our next talk will be Dr. D. Ieva. Thank you. Hello, my name is Antonio D. Ieva, and I'm professor of neurosurgery and the head of the computational neurosurgery lab at the Macquarie University in Sydney. Computational neurosurgery is a new field in which we apply computational modeling and artificial intelligence to diseases of neurosurgical interest. And here I would like to show you a workflow in which we can apply artificial intelligence from scratch. We're talking about neuroimaging, digital pathology, and also cognitive neuroscience study. So the first part is neuroimaging. We start from automatic segmentation of the microenvironment of the tumors. And you may know the challenge called the BRATS, in which we apply double learning for automatic segmentation of the edges of the tumor. Here you can see our contribution in which we had very high accuracy in automatically segment the brain tumor's edges. But moreover, we transferred this kind of learning to all the kind of brain tumors, including metastasis, vestibular schwannoma, and meningiomas. Moreover, more importantly, we were able to apply a new pre-processing technique by means of which we can use a much smaller data set than the BRATS uses. Why we use segmentation of brain tumors is because we like to do radiomics, which is the extraction and inference of features from the tumor in order to have a better diagnostic power. And this is our pipeline we built to have a better diagnostics, but even a prediction in terms of prognosis of our brain tumors patients. Translating all these neuroimaging things to the digital pathology, we can do what we call pathomics, which is exactly the feature extractions from radiomics, but applied to slides. This is an example in which we applied the double learning for automatic classification and the grading of glioma. And more importantly, we did also radiogenomics study, in which we were able and tried to predict the genetic mutation. For example, here I can tell you that no neuropathologist in the world can tell you if these gliomas are IDH mutant or wild type, even not, just using immunochemistry or genetic sequencing. So we thought to apply the double learning model by means of which we were able to segment and find that 85% of the people had 85% of the chance to have the mutation. The problem is that 85% is not good enough, is not a gold standard. So don't struggle enough or much trying to understand who these actors are, because these are synthetic actors generated by a double learning model called GAN. So by using the 400 glioma we had, we generated 3,000, 4,000 glioma as the algorithm is doing here. And by doing that, we had a very good training data set able to give us an accuracy of 95% on IDH mutation. So this means that many times, even from frozen section, 10 minutes for HND staining, and after that we can know intraoperatively if the patient has already an IDH mutation or not. Last but not least, it's very time consuming to have the ground truth of the images. So many times we have to do segment this metastasis in order to teach to the computer where the metastasis is. So my idea was that we use the eye tracking as a surrogate biomarker of a cognitive function. And by taking a radiologist and neurosurgeons looking at images, we record their scan path, which is the eye tracking with the movement of the eyes. And in this picture, you can see many neuroradiologists and neurosurgeons looking at images. And here is a summarize the pipeline we did to transfer the learning of an expert looking at the image to a machine learning model. And by doing that, we don't need to do the ground truth in a manual way anymore because the computer is doing the job as good as we do, but even much faster. So why we do all these fancy things is because this is the MDT of the present, but I don't foresee very far the moment in time when this little guy will sit at the table with us in order to help us to enhance our diagnostics and their decision making towards what we call the real AI of the future, which is augmented intelligence. So I would like to thank all the people working at my computational neurosurgery lab. And if there are interested in neurosurgery in the field, we also started a new computational neurosurgery fellowship in Sydney. Thank you very much. Thank you. Thank you. Next we'll have Dr. Baskin awarded the James Rutka Pediatric Brain Tumor Award. Well, thank you. It's a great privilege to receive this award and I'm very humbled by the name of the individual who sponsored it because it would take two lifetimes for me to achieve what he did. So what I wanna talk to you today is something quite novel and different. Many of you may or may not know who Otto Warburg was. Dr. Warburg won a Nobel Prize because he discovered that cancer shifted to anaerobic metabolism. Instead of using glycolysis, the electron transport chain in the mitochondria is the way that they get energy. And the reason for that is that it's cheaper. You get more out for what you put in. And now as a consequence of this, because the electrons are traveling down the electron transport chain, they put out reactive oxygen species. So inside the mitochondria of cancer, there's ROS, there's superoxide and hydrogen peroxide. In fact, at moderate doses, it's actually a mitogen because they've evolved to know that. So we came up with this kind of goofy idea. Would there be some way that we could take these electrons singing down the electron transport chain and knock them out of the chain and produce so much reactive oxygen species inside the mitochondria that it would be fatal. And there's a tremendous amount of biophysics involved in this. It's taking a singlet electron and taking it to a triplet electron. And we had a number of very smart people advising us with this. But in the end of the day, what we did is we took cultures of GBM cells and later DIPG, put them in dishes, and we designed this device to put these various types of oscillating magnetic fields through the cultures. Nothing happened. Very depressing. We pressed on, we pressed on, and finally we found that an oscillating field ramping up and down would do it. So on the panel on the left, you can see that we're staining for superoxide. And over time you see, and we have interesting time-lapse photography, these cells go up and up and up. And then suddenly at about 300 minutes, boom, Caspase-3 gets activated, and you can actually see these cells explode into apoptosis. This is about two years of work. So this particular study is a subset of our studies that we did with DIPG, or more correctly, DMG. We know that these oscillating fields causes mitochondria to generate more reactive oxygen species, and we know that beta-hydroxybutyrate actually causes mitochondrial hyperpolarization and elevates ROS further. So we said, well, first of all, could we get the effect at all in DIPG? And if we put in BHB, could we enhance it? So the first thing we did is we went to animal models, and these are tricky, because you have to inject these things into the brainstem. It took a little while to learn how to do this, but indeed, this is a standard DIPG cell line that's available. You can see that using the correct sequence, the panel on the right shows you the brainstem's full of tumor, that we could prolong survival in these animals almost twofold. And that was really striking, and particularly for DIPG, since nothing really works for it. And then we turned to some basic biology and we said, well, how powerful of this? So we did clonogenic assays, and we said, what happens if we just treat for four hours? What is that like? Well, we got the same kind of kill that you can get with two gray of radiation, which is fairly significant. It's about a 45, 55% kill, and it was just as effective as a dose of radiation. Then we said, okay, what about ROS generation? Well, if we measured ROS and we used the right sequences, we could increase the amount of reactive oxygen species threefold. We said, okay, what about enhancement with the BHB? Yeah, well, with BHB, it was fourfold more. So that's 12-fold increase in ROS to very, very toxic levels. So we did some clonogenic assays, and we tried various things to see that we could verify the hypothesis. Ketogenic diets elevate BHB, so there's a variety of things you could think about in terms of interventions, and what this slide shows you is that if you put BHB in the dish alone, nothing happens. You need some basic level of increase in reactive oxygen species with the oscillating fields, and then if you add BHB, boom, you knock it down further. So this was very, very exciting, and we, long story short, in four minutes, we applied to the FDA for compassionate use, and this is our first patient with DIPG. We've treated seven patients with GBM. Come tomorrow, and you'll hear about that. This is a 28-year-old woman with a HVK720M mutant brainstem tumor, had a suboccipital cranium biopsy, and it's proven, had radiation therapy, nothing worked, no chemo, no steroids, came to us desperate, and so about six months ago, we put this helmet on, and you look at this helmet, it looks like a bunch of beer cans on somebody's head, but these are three very powerful magnetics that are going off in sequence, producing quite a few millitesla through the brainstem, and you can see what happened. Over six months, you see the shrinkage of this contrast enhancement to the point where we're not sure there's anything at all, and if you look at the T2 flare, you see a similar reduction, just as we see in GBM patients, so this is like really exciting because DIPG is so miserable. So the take-home message is oncomagnetic therapy elevates mitochondrial ROS, kills DIPG cells in culture, kills it in animals, and maybe kills it in patients. You can enhance this with adding beta-hydroxybutyrate, it's a synergistic thing, and this is really a very promising and somewhat disruptive technology for treatment of DIPGBM, and by the way, other cancers as well. Now, this is not Optune, it's a helmet, you can take it on and off, you don't shave your head, it's only six hours a day. There's a completely different mechanism of action from Optune, we actually can get cell cycle arrest, there's no dermatologic side effects, and we see marked shrinkage. So I certainly want to thank the team, it takes a lot of very smart people, a lot smarter than a dumb brain surgeon like myself to figure out all this physics. This is really incredibly exciting and powerful and disruptive technology, I think it has great potential. Stay tuned, because you'll hear about clinical trials before too long, thank you. Our next speaker is Dr. Moore. Hello everyone, my name is Elyse Moore. I'm a post-baccalaureate research fellow at the NIH. I work in Dr. Prashant Chidibuyana's lab, which is the neurosurgical unit for pituitary and inheritable disease, and today I'm gonna be telling you about some of the work that I've been doing over the past two years, looking at how we can recapitulate the evolution of the pituitary gland and create new models and novel implantable devices. Before I get started, I want to acknowledge that this work has been funded by a competitive grant through NINDS and supported by Kevin Chen of the NINDS stem cell unit. So what are we doing and why are we doing it? Well, we know that patients with pituitary tumors make up about 10% to 30% of our population. These pituitary tumors can range in phenotypes, but in general they can cause a number of complications due to the cell type that's present or due to the treatment intervention that occurs. So many of these patients, one of the treatment problems that occurs is hypopituitarism, where the pituitary fails to secrete one or more of the hormones. This can be incredibly disruptive for the patient's life, especially if it is a long-term problem where the patient would have to go on hormone replacement therapy. Additionally, even though that these tumors are quite prevalent in our population, we don't really understand how they form, why they form in so many people, and the only models that we have to study them are really mouse cell lines or short-lived patient-derived cell lines. So to make up for this, there have been a few protocols that have been published over the last couple of years. The first you see here is for 2D cell culture differentiation from embryonic stem cells, and the second is for 3D organoid generation from embryonic stem cells. These programs are great and they do produce functional anterior pituitary cell types. However, they have a few downfalls that would preclude clinical translation because of implantation risk due to genetic manipulation. So for us, we want to improve upon these models and use human cells, but get rid of these translation precluding conditions and create models that could become clinical in the future. So with the protocol that we've developed over the past two years, we have been able to successfully differentiate corticotrophs in about 30 days from human-induced pluripotent stem cells. Induced pluripotent stem cells offer us the capacity to derive cells from patient samples, and as you can see here with this immunostaining, we are able to pick up both TBX19 and POMC, which are our two mature corticotroph markers. We can corroborate this data with bulk sequencing, where we see that our cells are progressing down a nice stepwise timeline from pluripotent stem cell to mature corticotroph and pituitary cell types. Though this doesn't tell us anything about the functionality, we then have to look towards, well, what are these cells secreting? Are they able to secrete ACTH like a normal corticotroph? To do that, we look at ELISA data, and this is showing us that both in a basal secretion level and a stimulated condition with CRH, these cells are able to appropriately respond to the endogenous stimulus that we know to stimulate the secretion of ACTH. So this is great, but what does it mean for you guys? I mean, you guys are neurosurgeons, and I'm a basic scientist, so you're probably wondering, how does this apply to you? Well, the first thing we can do with these cells is we can use them as a human model to study the pathogenesis in the gland. We can do this with genetic manipulation and further downstream study, but we can also move this into a 3D model and use it as an implantable device for those patients who are chronic hypopit and require lifelong hormone replacement therapy. We can give them a new option for, well, we can derive your own cells, differentiate corticotrophs for you, and then re-implant them to recapitulate the lost gland function. So with that, I want to thank the members of my lab, especially my mentor Prashant Chidibuina and Kevin Chen and the entire NINDS Surgical Neurology branch for their contributions to this work, and thank you for your attention. We're going to transition to the cellar portion of this talk, so just give me a minute. And our Dr. Lee, Dr. Jones, all right. So our first speaker in the portion of the talk entitled Cellar Tumor Challenges is Dr. Lee. Dr. Lee. Okay, thank you so much. So, I was tasked to—oh, sorry, this is the wrong one. Let's go back. There's another one for stellar tumors. Okay, perfect. I was tasked to discuss the limitations of endoscopic approaches. I don't have any disclosures. I think that when this type of talk happens, everybody likes to talk about the reach that you can have with endoscopic surgeries, show you things that we can't reach, and impress you with all the things that we can reach. I thought it would be interesting to take a different approach and talk about the limitations to endoscopic surgery that we often don't acknowledge, but are really major, which are the consistency of the tumor, the pathology of the tumor, if you need to have drainage after the surgery, and then a lot of the interoperative decision-making, which is really the difficulty here when deciding what to do and how much to do with these tumors. So, I'm going to go through some cases of mine to sort of illustrate some of these different things. This is a 16-year-old who came to me with vision loss and this heterogeneous sort of odd-looking lesion. We did a typical endonasal endoscopic approach to this tumor. Really odd, very sticky and ugly consistency, had some fluid components to it, and a really thick rind that had some movement that looked a little bit like it could be a craniopharyngeal type of a wall, but the pathology was very inconclusive and really odd, non-craniopharyngeal-looking tumor. Ultimately, had a very difficult diagnosis for the pathology, sent out many places and thought to be an ectopic salivary gland neoplasm, and at the time, you know, there's a lot of discussions between me and my ENT partner. We could expand and take the wall if we wanted to, which if we did and turned out great, we would say was the right thing to do, but ultimately we left the rind. Everything collapsed and he's been stable for four years. This next case is a 74-year-old male who came with vision loss and this cellular and supercellular lesion, and he underwent an endoscopic and a nasal approach. This tumor felt very much like a meningioma, although the pathology, both frozen and the permanent, was actually adenoma, but just really ugly, stuck to everything, bloody, really difficult. We tried sonoped, tried everything, and was able to core out the middle of it and get everything to collapse so that his vision could improve, but really was extremely limited by the consistency of the tumor. I know you've all felt the frustration of these and was just really limited by that consistency. This next patient is a gentleman, 45-year-old, who came with vision loss and this heterogeneous-looking tumor, which ended up being a pituitary adenoma with lots of septated pockets, which I think can be difficult as you work through each of them to make sure you're not going to leave residual tumor, and the capsule itself didn't collapse. This gentleman was unfortunately on immune modulators for vasculitis and things and ended up with a really rare complication of having Klebsiella infection within the cellular cavity, which we couldn't find a source for, but ultimately with some opening for drainage did quite well. This is a patient with an unusual-looking lesion, 67-year-old female who presented with vision loss. Really difficult to know what we're in for going in there and did a really small opening to start, just because we're trying to figure out what we had, and it was very evident as soon as we opened the cell, this was a Rathke's cleft cyst, and we're able to get all the cellular components of that, but unable to get the supercellular component of that from the approach that we had, and intraoperatively had to discuss if we create a CSF leak, we've created a major problem because then we have to wall it off, and so ultimately decided to just evacuate the cellular portion, hoping that it would evacuate the supercellular portion, which it didn't, unfortunately, but her vision did improve dramatically, and so I follow her chronically to decide if she ultimately needs further surgery. This is a 16-year-old who had a history of a craniopharyngioma resection as a child. All of his intracranial disease was cured, and he had this cellular disease that was left over. He had lots of headaches and was convinced that if we took out his tumor, we would cure his headaches. I assured him that was not the case, but that we could cure him from his disease. He could live lifelong without having any disease going forward, which is really important in a child. He's already panhypopit, so it felt like a freebie, although when we're in there intraoperatively, as you know, craniopharyngiomas are never a freebie, and it was stuck to everything, including the cavernous walls, and so I realized when we were in there, although we were able to get all the calcified pieces off of the cavernous walls, that telling him that we could cure him of his disease was maybe an overpromise, although we were able to do it, that there's a lot of pulling and tugging and saying to my partner, should we pull this? I don't know if we should pull that, that sort of thing. Ultimately, I was able to cure him of his disease and did cure his headaches, which I took credit for, even though probably had nothing to do with it. And then this is a patient with a really large pituitary macrodenoma, had cavernous extension, and a huge amount of disease in the clivus itself, which we see very often, and in patients in which the goal is to cure them of all their disease and a favorable consistency, and not in the cavernous sinus, we'll definitely take the time to drill out the clivus and make sure that we get all of that tumor. In this case, we did some of that, but just knowing that there was so much extension of this tumor, that there was no way that we could do that, we ultimately felt like that wasn't an appropriate thing to do. And then lastly, just a comment about cellar mets, a metastatic disease to the cella. A lot of times, these are weird-looking lesions, you're not sure what you're in for. As soon as you get in there, they're so sticky and ugly and bloody, and knowing that really early on to wait for the frozen and diagnose it, because ultimately, endoscopic surgery is not going to help this patient in the long term, as far as like a resection, identifying that can be so key. So in conclusion, I think that the reach is often what we talk about, but the consistency and pathology of the tumor really dictates the interoperative decision-making. And with that, I thank you so much for your attention, and I welcome any discussion after. Next, we have Dr. Roper. Thank you, and thank the organizers for allowing me to be here today. I'm going to talk about complication and avoidance in cellar surgery. Let's see. Thank you. I'm a consultant for Medtronic on an epilepsy clinical trial. So there are two components to complication avoidance, awareness of the risk, and then mitigation of the risk. And this is just a recent meta-analysis that lists some of the common complications of pituitary surgery. We could probably all kind of think of most of these things on our own. Thankfully, most of them are quite rare. But I'm going to talk about two in the limited amount of time that we have. These are two that over, you know, 30 years, I just can't make go away. And they are CSF leak and delayed hyponatremia. So I don't feel that it's possible to completely avoid intraoperative CSF leak with a pituitary adenoma resection. This is just data from our own center that's in a paper that we've been working on for too long. But anyway, the intraoperative rates of leak are here between 10 and 20 percent. We were comparing microscopic and endoscopic. It's not relevant to this talk. And then the one you want to try and reduce is the return to surgery for postoperative CSF leaks. And those are thankfully much smaller but still present. The reason I think that we have to deal with intraoperative leaks is partly because of this. You know, when I started doing these things, I had this idea that the diaphragm is this nice, sturdy layer of tissue with an opening just as wide as the infundibulum. And of course, that's true for some people and not true for others. Even back in, you know, 1975, Dr. Roten was saying in cadavers, about a little over half the cadavers had large openings in the diaphragm. And you can just imagine that if this is the cell that you were born with and then you develop a pituitary adenoma over the years, that the tumor is just going to grow right up through that hole. And so when you take it out, it's not going to be dura that you're up against. It's going to be arachnoid. And when we're taking out large tumors with just arachnoid, at some point, some of those patients are going to have a leak. This was echoed by another recent paper by Amadi et al., where they even proposed this ratio of the opening of the diaphragm compared to the area of the diaphragm and suggested that when that opening is big enough, that's why we see some of these tumors. Some of the tumors grow up to the brain and into the brain and some grow into the sphenoid. And that might well be a determinant of that or at least one determinant. So I do all these cases endoscopically with my partner, Dr. Brian Lobo of ENT, and I'm very lucky to have him as a partner. And now that we understand that it's really the repair that's the most important thing, I have to apologize because I don't do these repairs, but the, of course, the hallmark is the nasoceptal flap. And I think as was alluded to earlier, the important part of what he does is this pedicle sparing approach on all our patients so we don't have to know ahead of time whether we're going to get a leak. He makes this opening. This is the sphenoid ostium. And he makes this cut with preserving the posterior septal artery and then does his dissection and leaves that intact. So if we do get a CSF leak, it's relatively simple for me because I'm not doing it, but it's it's fairly straightforward for them to come to go go and create the flap and then do the repair at the end. And he's thankfully extremely good at that. So what is important for us to know as neurosurgeons probably is, you know, how do you take care of these patients after he's done this elegant repair? And this is our kind of routine. Five days of antibiotics, head to bed 30 degrees. In terms of mobility orders, 24 hours bed rest, 24 sitting, 24 walking, and then discharge. It usually turns into a four-day hospital stay just because the 24 hours don't hit at exactly the right time. They all get nasal saline, a Q2 hours while awake, and they get a post-op check with a scope by Dr. Lobo at two weeks. So that's how we manage that relatively common complication. The other one that I can't quite make go away is delayed symptomatic hyponatremia. This is another meta review recently. Incidence is somewhere between 4 and 20 percent. I'm sure incidence is somewhere in that number. This is something that happens with an EDIR of sodium levels usually about seven days post-op, which of course in almost all our cases these people are at home at this point. There's usually a decline over days four through seven, and this is from the literature, but certain risk factors that have been correlated with this problem include the presence of Cushing's disease, the rate of the sodium decline, and increasing age. Etiologies, we usually don't work this out at any specific level, but you probably usually ACCH, although cortisol changes and cerebral salt wasting may also play a role. So again, I've never been able to figure out how to make this go away, so the important thing is how do you detect it. This is the algorithm that we use. Everybody goes home on hydrocortisone either 20 and 10 or 15 and 5. At home, the patients get sodium levels every Monday and Thursday for two weeks in a pituitary hormone panel at two weeks, and then we see them back at either two or three weeks in our post-op clinic. The important thing here is actually having someone making sure they're getting the lab tests and someone who's checking the results, which you know this is once people leave the hospital, sometimes that's easy and sometimes it's not, but the important also part is that mild hyponatremia can often be managed via phone with just a voluntary fluid restriction and maybe some PO saline replacement if it's addressed quickly, and if not, we just have to tell the patients that they're going to have to come back into the hospital. If they do, this is a relatively straightforward problem to take care of. We usually just treat them with fluid restriction, IV saline, either 2% or 3%, depending on how severe it is, and endocrine consult, but again we never really end up with any specific answer, but over two or three days it usually goes away. The other important thing is it's almost always just a one-time event. It's extremely rare for it to recur once they're back at home. So this is my summary slide, but I'll end it there in the name of being brief. Thank you. Thank you, and our last speaker is Dr. DiNapoli, after which point we'll ask the three speakers to come up for a panel discussion. Afternoon, everyone. I thank the panel and the Society for the opportunity to speak. They asked me to talk about cellular tumor challenges, indications and limitations of radiosurgery. I have no disclosures other than the fact that I am a skull-based neurosurgeon and not a radiation oncologist, so just bear with me on that. So, you know, in thinking about it from a surgical perspective, how do we think about upfront treatment for considering radiosurgery and paracellar lesions versus surgical indications? And in our Center, some of the principal reasons we might think about something like that is smaller tumors, patients that have medical comorbidities that prevent them from surgery, or patients that simply just don't want to consider surgery. We certainly consider it more for meningiomas than we do for pituitary tumors because pituitary surgery has remained the mainstay for treatment in that disease. Also in adjuvant treatments, so for subtotal resections where tumors left in places where critical anatomic structures can't be avoided or it can't be resected, such as the cavernous sinus, it becomes a much more relevant treatment. And then also in the recurrent setting, when people have had prior surgery, it's also very useful. So in comparing, you know, radiosurgery versus radiotherapy, you know, what are the considerations in deciding between those two things? I kind of compared them side to side here, but, you know, greater than two millimeters to the optic nerve is required for radiosurgery for a single fraction, with less than 1% optic neuropathy risk. Abutting the optic nerve can be done with radiotherapy. Smaller tumors versus larger tumors. The rates of hypopituitarism are fairly comparable in the two diseases, or I'm sorry, the two treatment modalities. At roughly 10 years, it's about 25 to 30 percent, is what we quote our patients. The big risk with radiotherapy, obviously, and especially young patients, is the 2% risk of secondary tumor development. So that's not an insignificant risk for a very young patient. And then cognitive decline with radiotherapy, especially if you're treating large areas of the frontal lobes or temporal lobes, that can become an issue. And then wholesaler radiosurgery in a single fraction, because of the critical structure, such as the infedibular stock in the gland itself, can increase your risk for hypopituitarism. There was a recent published article we published with the IRRF Internet. It's a multi-cental consortium, kind of looking at dose constraints related to gamma knife radiosurgery and the critical structures in that area, and kind of came up with these constraints that we use to guide our treatment. So less than roughly 10 grade to the pituitary stock, 10 grade to the optic nerves in single fraction, 25 and 5 fraction, or 54 with radiotherapy, and less than 16 grade to the pituitary gland itself. So the first case I had today to show was a 64 year old gentleman who had had prior endoscopic endonasal surgery and was lost to follow-up. Initially presented in 2001 for that surgery, and then I saw him back several years later in 2021 with a recurrent tumor. The tumor showed optic compression. He was having some visual loss, visual field loss, and progressive headaches, and we elected to do an endoscopic endonasal resection for him, knowing that there was tumor within the right cavernous sinus that was probably unresectable, and then planned subsequently for post-operative gamma-niferative surgery to a treatment dose of 16 gray. So this is his pre-operative scan here. You can see on the right side, the tumor does involve a decent amount of the right side of the cavernous sinus. And then this is our post-operative scan. As expected, the tumor there on the right side of the cavernous sinus remains around the carotid, sort of lateral to the carotid, where the ocular motor nerves would reside. So we don't necessarily want to go digging in there. And then post-operatively, the tumor volume ended up being six cc's, roughly, to a prescription dose of 16 gray. And the coverage ended up being 99.8% in a single fraction. So that was a good treatment for him. We kept the optic nerve dose less than eight, which is, we can go to 10 in certain situations, but we like to stay less than eight at a one millimeter margin. And that was still 8.2 at a one millimeter margin. So he did really well, and his tumors remained stable. So we actually got some shrinkage of his tumor at the six month scan, approximately three millimeters. The second case I had was a clinoidal meningioma. This is a 56-year-old gentleman that presented with, or I'm sorry, woman that presented with sensitivity overlying the skin, jaw pain. And it revealed a left clinoidal meningioma. The patient was followed expectantly for a while, and then the MRI revealed interval enlargement of the tumor. The visual exam was normal, no visual deficits, and the hormones were also within normal limits. You can see the dimensions of the tumor there, and we just discussed with her surgery versus radiation options. She was really reluctant to undergo surgery. There was also some things about the tumor that kind of pushed us towards a radiation option. The tumor was kind of 280 degrees, or at least 180 degrees around the carotid there, and also kind of sneaking into the medial aspect of the optic canal. So she didn't want to take that risk and thought that was reasonable. So we opted to treat with five-fraction Trubeam-Linek therapy to a total dose of 25 gray. So we did a pre-plan on the gamma knife to see if a single fraction was feasible, but with the abutment of the nerve, we had about a millimeter separation, and it wasn't just feasible to keep the dose less than 10. So we treat to a dose of 25, the optic nerve with five-fraction tolerance is less than 25, so you can cover it. We covered 99.8% of the tumor without any issue with the optic nerve. And the single-fraction dose would have resulted in a pretty high dose of the nerve. So the last tumor I was gonna talk about is a 39-year-old woman that presented with a very large secreting growth hormone tumor. She had acromegaly. We elected to undergo, and she had elected to undergo endoscopic resection. And she had a residual tumor in both cavernous sinus, which was expected pre-op, because she had very lateral extension of the tumor. And she remained, her IGF-1 levels remained elevated after surgery as well, even on medication. So we discussed with her the option of radiotherapy for this tumor. She complained of retroverbal headaches and double vision as well. So this is her, let's skip that one. This is her preoperative scan here, just kind of showing a large expanded cella extending into the supercellar space encasing both carotids well into the lateral cavernous sinus compartments there with the pituitary gland pushed downward. This is her postoperative scan after the endoscopic resection. So we got a pretty good resection. We decompressed the optic nerves, significant debulking of the tumor. Got a good amount of the tumor out of the left cavernous area but the right cavernous sinus, as Karen was talking about, was very sticky and bloody and thick. So we weren't able to asafely reduce that out of the cavernous sinus. And she opted to undergo a radiotherapy plan for 54 gray to the residual tumor. Kind of shows the plan there in the middle. So a total dose of 54, like I said, with 52 to the optic nerves. And ultimately she did really well. That's her preoperative scan there on the left. And after her surgery and the radiotherapy plan, her tumor is grossly on the MRI. You can't even see that it's there. So had a great result. And her IGF-1 levels did normalize to high normal after the radiation. So kind of in summary, just kind of summarize the talk. SRI surprised high tumor control rates, greater than 90% with acceptable rates of newer worsening endocrinopathy as well as optic neuropathy. The principal limitations for gamma knife at our center is proximity to the optic nerve. You need at least two millimeters of optic separation. Non-functional adenoma status, younger age, higher margin dose and higher doses of the pituitary stock and normal gland. We're independent predictors of newer worsening hypopituitarism. And we still think surgery remains the mainstay of treatment in parasolar lesions if the mass can be safely approached and resected. But when resection is limited by critical structures, it's a very effective adjuvant or even upfront therapy. Thanks. Let's stay up here. Why don't we open it up for questions for our panelists. Is Dr. Roper still around? There he is. I'll scooch out of the way. So these are great talks. I really enjoyed listening to them and appreciate the subtleties of the surgery. Having done a fair number of myself for over 30 years, I have a couple of comments, things that I've learned. In regards to your talk, I have really learned not to be so afraid of a transplanum approach. I used to be scared to death of it. But a lot of the problems you talk about, if you can just drill off a little bit of the planum sphenoidale, you can go over and on top and you can dissect the A2s off. And it's amazing how much more access you can get. Of course, as was said, you have to have, either know about nasal septal flaps or have someone who can do it. Although there is the inferior turbinate flap that you can harvest, and you can also harvest the fasciae latae flap. So that has really expanded our abilities to get some of these sticky tumors out. The second is not to be so afraid of the medial cavernous sinus. That's something I was taught to never do. But really, it's not a one vein. It's little compartments. And so you can open the medial cavernous sinus. And we have some people talked about this here at this meeting, incredible stuff. I don't go past the medial. I mean, I stay medial to the carotid. But you can do quite a bit in the cavernous sinuses. As far as hyponatremia, we have a protocol where we put patients on fluid restriction 1,000 cc's a day for a minimum of five days. And our readmission rate for hyponatremia dropped by about 60%. And so everybody goes out on 1,000 cc's a day. And it really has made a tremendous difference. The points made, you have to get good follow up. You have to make sure they get the blood draw and somebody actually follows up on it. You know, they go somewhere else and they don't send it in. But just that fluid restriction for five days has made a big difference in our readmission rate. It really was about 20, 25%. Now it's about 5%. And the radiosurgery talk was great. You know, fractionation has made a big difference. It's, we've also had the experience that using just the gamma knife is too risky, as great as the gamma knife is. So you can either do five times five or the old protocol, which worked very well. We published this was, what, 54 gray at 1.8 gray per fraction. The optic nerve tolerance is very good. So these were terrific talks that bring out a lot of points. And these are things I've had to learn the hard way. So I thought I would just share those few items. Great, other comments? Molly. Yeah, hi, that was, those were all wonderful talks. For Dr. Roper, that was great. And I agree with those thoughts on the delay of the, or treatment of the post-operative hyponatremia. Do you have any thoughts about the other, you know, the literature's coming out to support that? I think Dr. Q, who's in the audience, has a nice review, most recently, about post-operative fluid restriction in a scheduled way. The other thoughts are, what do you think about a scheduled sodium check in the post-operative course? For us at our shop with three pituitary surgeons, we all have the same protocol. They go home on a liter and a half fluid restriction from day four to 12, and they all return, or they have to get lab-strawn at day five or six. With the sodium check, we oftentimes do ACTH and cortisol also. I know there's limitations, but what are your thoughts on that? I hadn't thought of it. I mean, so you guys are sending them out, everybody goes out, and they're actually doing it? 1,000 cc's a day? We actually, wow. My patients wouldn't. We send them home with a bottle that's about a liter, liter and a half, and we say, and it's got some branding on their tooth. This is L.A., so we gotta do a little marketing, but we want you to drink one and a half of these a day, and it's a surprisingly good adherence to the protocol. That's great, but you still readmit a few? No, so in the last two and a half years since doing that, both sites, not a single readmission, and the reason, I think, is because I try to get them back personally into clinic. When you force them to come back, it forces them to take responsibility, is my thought, and you can get to these things before they dip into it. They're gonna have to report to the principal's office at some point. I've been accused of being a principal. Well, I think it's a great idea. For me, it's only a couple or three patients a year, so I guess I haven't thought about treating everyone for it so what we're trying to do is, and again, it's only partially successful. We're just checking it two times a week, but they miss it. It gets done in the panhandle of Florida, and three days later, we get it back at 120, so I think it's a great idea. Thank you. Thank you. Other thoughts? You know, we tried the bottle thing, too, like Same Home, and we've had a couple patients bring back the pitcher filled to that line with urine. So we've had to be really explicit in our instructions. All right, well, thank you, everybody, for your attention. I thought that was a great session. If there's no other questions or comments from our panelists, I guess maybe we'll end a few minutes early. Oh, do you have cases? Do you wanna do that, or? Yeah, let's do that then. I just had two cases. I didn't do anything. My clock just came on. There we go. Yep. So I just had a couple interesting cases to present for kind of parasitic lesions that maybe we hadn't discussed yet. So my first patient was a 55-year-old gentleman presented to the ED with new onset seizure, left-sided weakness. His MRI revealed a large homogeneously enhancing mass consistent with the clinoidal slash sphenoid wing meningioma. He had a lot of peritumoral edema, mass effect. He had some cognitive changes. He was a pretty high-functioning guy, financial advisor. So he was taken for resection, and his postoperative scan was gross total resection without residual tumor, but his pathology was consistent with grade two meningioma with nine mitoses per high-powered field. Here's his initial preoperative scan. So pretty large tumor, a lot of mass effect, brain compression. It was kind of sneaking out the optic canal, so we did an optic nerve decompression as well. And this is his postoperative scan, which looked good. His initial postoperative scan on the left there, we combed through it pretty closely and couldn't really see any postoperative enhancement that the radiologists could find, but over time, at his 18-month scan there, you can see just medial to the clinoid, the carotid there, kind of sneaking into the clival recess. There's some recurring enhancement. So I think it's kind of an interesting discussion. I don't know what people around the country are doing, but at our center, we've kind of adopted kind of a watch and wait philosophy with these patients as long as their scan looks, we can be convinced that their scan is clean after surgery, even if it is grade two. Obviously, the downside of doing upfront radiation is that you have to radiate the entire cavity and expose a large portion of their brain to radiotherapy. So in a young, high-functioning patient that could have a lot of long-term sequelae, so we ended up treating that small recurrence there. It's a sixth grade with five fractions, so a total dose of 30, and he's been stable since that time. So as I said, I kind of just debate in the literature, maybe some of the ongoing trials and some of the talks we heard today, molecular studies may answer this question for us, but do we treat grade two upfront with radiation, or do we wait and watch and try to treat just the small recurrence? And sometimes you can treat that with a single fraction, and that's the best for the patient, so maybe we'll figure that out soon. Sure. I think I would have done exactly the same, and I think age is so important when deciding that, but I would say one factor that we don't talk a lot about is that if a patient, if you really believe they're gonna follow up or not, because we have a lot of patients who might be homeless or socioeconomically disadvantaged, and sometimes if you lose them to follow up and they come back with the same giant tumor, that's just such a massive problem, and so I do think it's important to think about, are they gonna follow up, and is that a factor in whether we should radiate upfront? For sure. Yeah, I think that's a great point, especially in some of the more rural areas, we tend to treat them upfront. The second patient I had was a little more straightforward, I guess, a 53-year-old right-handed woman presented with slowly progressive visual field loss, the right temporal hemifield. MRI revealed a large enhancing, moderate-sized enhancing mass in the tuberculum cell area, consistent with meningioma with optic nerve compression. She was taken in section, her pathology was also a grade two. So this is her initial preoperative scan there, you can kind of see the tumor. It is contained mostly in the midline between the carotids, kind of between the goalposts there, moderate-sized tumor, so I guess the considerations with this case, which would be interesting for discussion, is open versus endoscopic approach. I think either could be entertained. We elected for an open, mini-tereonal approach, just because of the progressive visual field deficit. We really wanna make sure we can decompress her right optic nerve well, and make sure we got the tumor away from it. So here's her postoperative scan, the three years. Looks good, and she has had no recurrence to date. We ended up not radiating her as well, even though it was grade two, so it's kind of a similar discussion, but we've tended to, with these kind of tumors in the tuberculum cell with optic nerve decompression, we've tended to do them open, but I know some people will do the optic nerve decompression from below, and you can get a pretty good decompression with that. We did a lab study once, and kind of looked at the degree of decompression. You can get a nice, at least 180-degree decompression, if not 270, so. Can you put the images back? Sure. And then maybe just out of curiosity, a show of hands, who would do this endoscopically? And open. Something we're not gonna do. Yeah. I think you can go either way. I have a comment about this, again. I think you can't argue with the result, it's terrific, but one of the things about these tumors that I've found is it's, if you come in terionally, it's the contralateral optic nerve is very easy to decompress. The ipsilateral optic nerve, you're in the optical carotid triangle, it's very narrow, you try to come underneath, and sometimes you manipulate the ipsilateral optic nerve more than you want to. I agree, yeah. And so I've had some patients with diminished vision, and when you go transplant him, if there's not too much lateral extension, it coming underneath the chiasm and peeling it off, you're much more gentle, and I, at least in my hands, the risk of optic nerve injury is lower, transplant him. But I mean, you have great operation, great result. Can't fault it, but I just find that that ipsilateral optic nerve sometimes is really a challenge. Yeah, especially if it's really, it's really tented. You're trying to come underneath, you can't get there. I've tried, yeah, we try to reduce it from other without kind of manipulating the nerve itself, but yeah, I agree, it can be challenging, and I've done it both ways. You know, the endoscopic approach does definitely, I think, take longer for the patient, and you have the risk of CSF leak for bigger tumors. I think this one, you could have done either way, honestly. It's good for endoscopic, it's right in the midline. Those are the two cases I had. All righty. Well, thank you, everybody, for attending and for your attention, and have a great day. Have a great rest of your conference. Travel safe.

brain tumor

spinal metastasis

quality of life

artificial intelligence

financial burden

cancer patients

spinal metastasis management

intraoperative imaging techniques

brain tumor surgery

personalized approaches

multidisciplinary approaches

malignant meningioma

KL4 gene

oncogene

tumor suppressor