Thank you for attending today's Ask the Expert webinar on Making Sense of the Alphabet Soup-Practical Tips on Genetic Testing in Pediatric and Adult Patients with Epilepsy. Please note that this webinar does offer continuing education credit after completing the evaluation. You will be taken to the evaluation immediately after the webinar. You will receive an email when the on-demand version is available. Before we get started, I would like to take a few moments to acquaint you with a few features of this web event technology. At any time, you may adjust your audio using any computer volume settings that you may have. On the right-hand side of your screen, you will see the text chat window. There is a large window which holds all your sent messages and a smaller text box at the bottom where you will type in your questions. To send a question, click in the text box and type your text. When finished, click the send button. All questions that you submit are only seen by today's presenters. Your questions will be answered at the end of the webinar. Any handouts associated with today's webinar can be found in the handout pod on your screen and under the resources tab on the event page where you first registered. I'd like to introduce today's moderator, Dr. Wolfgang Molhoffer, AES Online Education Vice Chair. Dr. Molhoffer? Thank you so much for the introduction and hello everyone from all across the U.S. as I can tell by some of the names that I recognize. It's my pleasure to welcome everyone for today's live webinar. I wanted to get things started by mentioning that neither I nor Dr. Paduri have anything to disclose that is relevant to today's topic. I wanted to introduce our today's speaker who is Dr. Anne Paduri and I'm pretty sure a lot of people know her. She is a professor of neurology at Harvard Medical School and director of the Neurogenetics and Epilepsy Genetics Program at Boston Children's Hospital where she serves as Associate Chief for Academic Development in the Department of Neurology and holds the Diamond Blackfin Chair in Neuroscience. As a physician scientist with a focus on epilepsy genetics, her goals include contributing to the genetic landscape of epilepsy, creating models of human epilepsy, and developing novel treatments for genetic epilepsy. Upon completion of today's webinar, everyone should be or will be able to identify patients that might benefit from genetic testing, will be able to determine what types of genetic changes can potentially explain specific epilepsy syndromes, and then be hopefully able to choose the appropriate type of genetic testing for individuals with suspected genetic epilepsy. As mentioned previously, please feel free to enter your questions in the Q&A box to the right-hand side as the presentation goes on and certainly at the end of the presentation. And I will do my very best to summarize questions for similar topics to have Dr. Paduri answer them in the order received. And now I would like to turn it over to Dr. Paduri. Thank you so much. Thanks so much, Dr. Mulhofer. I'm really pleased to be here. I do recognize some names, so hello to those of you I know. And I'm looking forward to spending this time together, and I hope we'll meet those learning objectives. Hold me to it. If I don't, then I can fill it in at the end. So I'm really going to focus today on this introductory part about really thinking about why our patients have epilepsy. We know epilepsy is common. We know that the lifetime prevalence is 1 in 26. We know that you can start as an infant, as a middle or high school-aged child, or anywhere in between. But why do our patients have epilepsy? That's an important medical question. It's an important question for any family or for a person with epilepsy. And it's also really important for those of us who treat patients with epilepsy, individuals with epilepsy, largely because we have not only a medical and scientific interest in knowing why, but also because our treatments are empiric. And this was the case when I was in medical school. This was the case through my residency and fellowship in epilepsy. And you get to the end of several years of training and you sort of think, okay, well, the last several years of science must have answered a lot of these questions. We must be doing better. And when we look as a field at the drugs that are available, there are more drugs and maybe the side effect profiles are a little bit more tolerable in some cases, but our outcomes are the same. And one in three of the individuals we treat with epilepsy do not respond to the medications we have available. And that hasn't changed. So we're sitting here with this scientific and medical question of why do our patients have epilepsy? And then we're sitting here with this real clinical urgency, which is the medications continue to not work in one in three individuals and in the two-thirds in whom the medications do work, they're often not without a cost in terms of side effects. So there's a real urgency in my mind to try to figure out why, because if we can understand why in individuals and as a group, many individuals have epilepsy, we can then try to target our therapies, try to focus our treatment paradigms towards the individual biological causes. So that's really sort of the big picture why. We have had clues from our patients over the years. We've had clues in the form of MRIs. And so this was certainly something that in the last couple of decades went from we'll get an MRI for somebody with focal epilepsy as soon as we can to, you know, you really shouldn't be treating somebody with focal epilepsy with possible exception of some of the childhood epilepsy syndromes. But if someone has new focal epilepsy, we're going to want an imaging study. And even if they have a focal epilepsy syndrome, like what we used to call BEX and what's now the childhood epilepsy with centretemporal spikes, that, you know, still if there are features such as asymmetries or it's all one side or there's convulsive seizures and you're not quite sure how it began, we're still going to image there. So I'm belaboring this a little bit because of my age, because when we started, we had to sort of justify every MRI. But I think it's worth thinking about because I'm still following some of these patients from that era. And sometimes those clues were huge. This is an example of an MRI of a little baby who had relentless seizures, didn't respond to any medications. And here you've got a huge malformation called hemimegalencephaly. That sort of pointed us toward, first of all, a prognosis, which is this is going to be severe unless you do a resection. And also, you know, really the ability to move from medical to surgical treatment. That's fine. I really wanted to know why that happened, right? Why did this brain malform? And that actually took me from what I thought was going to be a full-on clinical career to a little bit of a left turn into a research project that turned into the rest of my physician scientist career. That's to sort of look at the brain tissue and take the DNA from the brain tissue. And I'm not going to talk about that a whole lot, but I mention it here because certainly how many of us were thinking about genetic causes of epilepsy even many years ago was that it could be sporadic and it could be due to a genetic change that was just present in a child's brain. So that's a big lesion, pretty rare. In a big, busy pediatric hospital, you might see something like this a couple of times a year versus the much smaller lesions we might see in individuals who have something like this, which is a subcortical, actually periventricular and transmantle lesion over here. You also see focal cortical dysplasias, which are much, much more common. Any epilepsy center will see individuals with epilepsy with focal cortical dysplasias every week, several times a week. And so they go from the large and very rare to these smaller and more common types of lesions with different effects. So that's accounting for a good number of our patients, maybe a third or so of them. But then we have the much more common situation where we have a child who has epilepsy who comes to us. And this example is the child who was age three when she started having generalized tonic-clonic seizures and then having many, many drop seizures per week. And the clue did not come from her MRI. Her MRI actually initially was read out as normal and continues to have normal cortex. The clue for her actually came from a family history where her older brother had had something very similar, seemingly normal development and then a regression in the setting of multiple generalized seizure types. It was really catastrophic for the brother. The family sought an answer in genetics because the family history was obviously pointing to that. They'd been told after the first child, well, this is something sporadic. And then when the second child came with the same symptoms, they said, okay, hold on, wait a minute. Maybe we don't know something that we should know. They actually came to me in a sort of roundabout way. And I tell this story only because I think that when we go to meetings and we, particularly now that we're more in person, we go to lectures and we have formal didactics and so on. That's great. But it's the chatting in between that actually leads to a lot of these connections. So a nurse who was at an epilepsy foundation lecture where I happened to be giving a talk, an old friend had asked me to give, came up afterwards and said, I have this family I want you to see. And they came to me with a variant and a gene called Grin2A. And if you know what Grin2A typically gives you, it can be focal epilepsy, it can give you a regression in speech, it can give you a number of different things. It doesn't usually give you generalized tonic-clonic seizures and drop seizures and those mixed seizure types. And it's also not typically in two siblings. And in this case, it was inherited from the parent. And then we said, well, that doesn't seem like the answer, does it? It's an epilepsy gene, but this doesn't seem like the answer for this family. And sort of that, knowing the phenotype that the children had, knowing the phenotype that that gene ought to be associated with was key to sort of putting together the answer that this wasn't the answer, that we should keep looking and re-analysis of the exome. Fast forward to a paper that came out of Austria within weeks of my seeing the patient led to discovery of a treatable metabolic disorder, uridine deficiency. And that allowed us to go from a regressive picture to actually a picture of recovery. After seizure freedom, still needs some medication to be treated, but a complete return of tone, motor skills, running, walking, running, swimming, and then eventually reading and acquiring cognitive skills. Not neurotypical, but still able to be schooled in a typical setting with a lot of supports. And a huge win in terms of genetics driving the treatment. In this case, in an oral form, it didn't require designing a new therapy. It was actually a, uridine was treated in other settings, safely in children for other metabolic disorders. So we knew it was safe. And I bring this up as an example of precision medicine that doesn't require, you know, millions of dollars and an army of people to put together, but just to say, you know, if you look carefully and you put the phenotype together with the genotype and it fits well, you might actually be able to do something for your patients that's very, very specific and very precise. Now, before I go on, I'll just admit people always say, well, that's fine, but how often does that happen? Not often enough. You know, that's one out of a thousand or so cases where you'll come upon something and you'll have an oral safe treatment you could just sort of give them. In this case, there's also a blood phenotype that we could use as a biomarker to see if we were dosing enough because we otherwise weren't sure if we were getting enough into the brain in addition to seizure reduction. So it's too few, but I would argue that we could make it more than that if we consistently look for patterns in our patients and consistently do the type of sequencing we're going to talk about. So what's the evidence that it's going to matter? We've had a unique opportunity through an institutional platform called the Children's Rare Disease Cohorts Initiative here in Boston, where we were asked to pilot sequencing in patients with epilepsy, along with our colleagues in gastroenterology who were asked to pilot their patients with inflammatory bowel disease. And this was a pilot that we started, we launched this about five years ago before the pandemic had started. And at that time at our institution, we could not do exome sequencing. We weren't allowed to send out exome sequencing. I will say some genetics colleagues would sort of get samples and sneak them over to a FedEx box and send them directly to various testing companies. I don't advise breaking all your hospital rules to do things like that. I think it's a little bit more effective and more straightforward to work within our systems and be able to offer to all of our patients the right approach. But we were asked if we would pilot this and the hospital agreed to pay for a limited number of cases to be sequenced in exchange for our making the data available to the entire hospital system and to help build systems to integrate genotype and phenotype, which we're still working on. That all sort of got built up and then we were doing blood samples and everything came to a grinding halt in March of 2020 and it gave us the opportunity to push the system a little bit, to push the IRB to say, well, why can't we do remote consenting? Like people have internet, right? So we actually shifted to a remote consenting model and we shifted to a cheek swab DNA sample model where the samples were sent directly to the laboratory, which was able to provide research grade sequencing data back to us, but also held a clinical sample for us in case we wanted to do a clinical confirmation. Very different than the genetic research of my postdoctoral era where people enrolled in a database, they enrolled in a study, they sent you a DNA sample and they never heard from us again. That might be fine for our colleagues who are purely researchers, but for those of us who look these families in the eye every time we see them, it's not good enough. We want to be able to give the answers back to the family. So we said to the hospital, we'll do this, but there has to be a clinical loop back to the families and they shouldn't have to come back and give another blood sample. They've got to be able to get this done. So that's how we did this. I give you this example and I'll give you the reference mostly because you can now use this if you want to push your institution to do something similar or to advise them to do something similar. It turned out to be great for us and for the gastroenterology group who did it. We now have over 40 cohorts in other disease areas that are adopting the same model. Along the way, it's been a little bit easier to get clinical sequencing done, so we do that when we can and we use these slots for those who are not able to get it done. So what did we find? We had over 500 kids with unexplained epilepsy. There are 500 little people here. A hundred of them actually had new answers from exomes and we were able to get those to patients up here. Another hundred had candidate genes, meaning we found really compelling changes, but they weren't in genes we knew about, they weren't in genes that anybody had heard about or recognized enough. When we went to PubMed, it sounded really interesting. Wow, these are genes expressed in the brain and maybe they lead to hyperexcitability, but it's not our household SCN1A, KCNQ2. It sounded interesting, but wasn't quite proven yet. So we're sort of in a holding pattern there and we're trying to model them in Zebrafish in my lab and mouse in other labs and in other systems to be able to try to figure out what they have. Then we have a whole bunch of these other kids, 300 plus other kids, where we don't have answers. And we sort of said, well, we'd love to move to genome and actually Alyssa Degama, who's on the call, I see, is looking at actually the genomes of the youngest kids there whom we didn't solve because it's not everybody who's going to get an answer. Now this is all comers. This is people with childhood onset epilepsy. Some of them were children when we enrolled them. Some of them were children when they had onset and they were adults when we enrolled them. And about 20% we had an answer that was able to be returned clinically. So keep that in mind because that's the exome yield for childhood onset epilepsy, any age, any range of intellectual disability or not, and so on. I'm going to tell you now about an international partnership where we got data on a small subgroup of children with epilepsy. So these children will grow up to be adults with epilepsy. So if you're an adult neurologist, don't tune out because this is relevant to your adult patients who had early onset epilepsy. What we did is we worked with our colleagues in Toronto, Great Ormond Street in London and Melbourne, Australia, to be able to accelerate the acquisition of enough data to make a case for rapid genome in infants. So what we did is we started at each of our institutions. We had parallel sort of a federated process of enrolling patients with childhood onset epilepsy, new onset with onset at less than a year of age and without a clear cause. So this was non-acquired epilepsy. We didn't know why they had it. People were inclined on the treatment team to pursue clinical testing. What we did is we offered them rapid trio genomes. So in our first hundred cases, we actually had nearly half rapid answers by putting together these cases. And this was within the span of a year. And this was work that the exome work was put together by Hien-Yen Koh and Lacey Smith and Alyssa put together very nicely that are these hundred genomes and we've published references here. I'm putting this out here because I'm very proud of the data, but also because this is sort of the practical on the ground demonstration that we wanted to do to show that all of this research had happened through many, many, many individual labs by consortia in Europe, consortia in the U.S. and Australia and all over the world to put some genes on the map, like SCN1A, which even when I was in med school was beginning to be talked about and published by the time I was a resident 20 plus years ago, but also a lot of newer genes, which we'd never heard of before. Lots and lots of research, but it wasn't in the, it wasn't in the, I'm seeing there's interference. Sorry. Let me know if I should change something. It wasn't in our abilities to do testing until more recently, so we really wanted to show that it was okay. All right, so what did we find with these babies? So this is what Alyssa put together and what we found was a whole host of genes. I'm just going to walk you through here from the kids who presented in the first few days of life to the kids who presented later in infancy and what you can see is a number of different things. One is that we have different epilepsy syndromes. The ILAE has nicely updated the syndromes and what we have is now several different syndromes including some self-limited neonatal epilepsies, but also some more severe infantile epilepsies and epileptic encephalopathies or developmental and epileptic encephalopathies that might present in infancy like the early infantile developmental encephalopathies and also infantile spasms, which are really kind of the more common of the rare early onset epilepsy syndromes. One in a couple of thousand babies will have infantile spasms, which is more common than I would have thought from having lived life until my child neurology residency and then sort of starting to see them, but it's certainly it's about as common as Down syndrome and that maybe is borne out in terms of people's practice in terms of how often you see it, but not always diagnosed. And so here's what we found, a whole range of genes, some old friends, so KCNQ2, SCN2A, you'll see in these early onset groups, so here's KCNQ2, SCN2A, but you'll also see those genes showing up later. So here's SCN2A later on, later on with another syndrome, which was self-limited. You'll also see some really nice prognosis genes, PRRT2. You get a child with convulsive seizures, it's pretty scary, and you find rapidly that they have a de novo or even inherited variant in PRRT2, and you say, well, we know that the prognosis is going to be good for this one. Maybe you'll look for dyskinesias, maybe you'll look for other things, but you can treat it with carbamazepine. It's not precision in the sense that it's based on the molecular structure or the molecular defect, but it is a gene-advised treatment that you might not otherwise use in an infant because it's hard to maintain carbamazepine levels, and a great prognosis. Others not-so-great prognosis, so some of these earlier onset severe developmental encephalopathies, like these metabolic disorders, the MOGS, for example, or the recessive BRAT1 variants, these are ones where the prognosis was quite terrible. These are very, very early on, so you can see in the first month of life, and with this early onset picture and a severe recessive disorder, we have genetic counseling, we can provide the family, but also we can provide them some direction, and unfortunately, bleak prognosis, which might allow the family and the treating team to make decisions about how aggressive they want their medical measures to be. So we learned a lot. Not new genes. We weren't here just to find new genes, but we're really here to characterize if you have a prospectively ascertained cohort, whether they're in hospital in the NICU or the PICU or the floor, or they're in your office coming with infantile spasms, if you relatively quickly get on this prospectively, sequence them right away. We didn't interfere. If people wanted to do a micro, go ahead and do what you're going to do, but we said, let's do this rapidly, get an answer back to these families and treating providers within a week or so, and now we're going to follow their developmental outcome and their seizure outcomes to see how much of a longer-term difference we could make. So our yield, again, was 43%. The impact was that nearly all of those who had a genetic diagnosis had a change in prognosis or a change in medical management in terms of medication changes or additional laboratory studies to do, as well as all of them having had genetic counseling regarding recurrence risk, which is particularly important for young families. So again, it's more than the 20% we saw in the all-comers, all-age group of onset. This is the first year of life onset, more than double the chance of finding something that's meaningful and that makes a difference for treatment. So I'm just going to scoot out for a second, zoom out a little bit, and just sort of talk to you a little bit about how we're approaching this big picture. This internal purple circle here is how I describe my Tuesday afternoon. I see patients who have a diagnosis of epilepsy, and of course, we want to confirm that diagnosis recurrent unprovoked seizures or the tendency to have recurrent unprovoked seizures. And then what do we do after we make that diagnosis? We do a genetic evaluation if we can, and then we, as we talked about earlier, have this empiric approach to treatment that really isn't informed for the most part by the genetic evaluation in most settings, especially. What are we trying to do? We're trying to learn from the genetic diagnoses or even the syndrome diagnoses by doing natural history of known genetic syndromes, or even if it's not known to be genetic, by those known syndromes that we put together and our ILAE colleagues to define some of those. And we continue with gene discovery. We then look at the variants very carefully to see if they're pathogenic or likely pathogenic, meaning, is this a variant that could be an explanation for this person or not? I mentioned the GRIN2A variants in the pair of siblings with the little girl who had regressed and then was improved with the iridine. That variant sort of looks, you know, maybe like it was doing something. Maybe it was affecting the protein, but it was not a good fit with the phenotype, and that's part of what goes into variant interpretation. It's really, really important to know that, that there's a certain amount of population data we can look at. If I, let's say I had epilepsy that was, and I had a variant in the gene DEPTIC5, which is a big gene. And you said, well, that's a focal epilepsy gene, so why not? Maybe that's why you have epilepsy. And you looked at a database, a population database, such as the NOMAD database that the Broad Institute has, and you said, well, one in five people in Europe have that, and we're not sure how many in South India where your family's from, Ann, but one in five Europeans have this. Do one in five Europeans have DEPTIC5 focal epilepsy? No, we know the answer is no, right? We know some of the epidemiology, absolutely not. That's gonna call that variant into question versus, oh, actually, there's nobody in this database of 120,000 people who have that variant. That seems like that's a rare variant that could be the cause of your epilepsy. So all of that goes into it, knowing what's in the population, knowing what the person has, knowing the phenotype. And so there's the review that we do one-on-one with our patients, and then there's expert variants review through ClinGen and ClinVar to sort of see which genes are really epilepsy genes, which variants are really compelling pathogenic variants. Okay, so in terms of what do we do next with the genes and variants we find, we have a functional assessment in cells, in fish, in mice, and so on and so forth. And after that, what do we do? In some cases, we can do preclinical trials that could hopefully lead us to precision therapies. So in that case, how are we taking all of this into account? How can we bring this to our patients? Is it really the hope is to bring all of these puzzle pieces from the outside into our patients and into the clinical experience? So I'm gonna talk to you a little bit about gene discovery and how far we've come. I mentioned that when I was in training, we had like almost no genes. I was asked a little over 10 years ago by Dan Lowenstein if I would review with him the genetics of epilepsy. He said, let's focus on the early-onset epilepsy genes since that's where a lot of the action is, but knowing that many of these early-onset epilepsy genes are also relevant to adults. And here you have it. This was all of them. It took like a Sunday afternoon to put them together because we really didn't have a lot of genes that we knew caused epilepsy. Now you might be looking at some of these and saying, oh, there's that sodium channel cluster. And yeah, I know about those sodium channel genes. You might be looking at some and saying, oh, KCNA1, what's that? And that has epilepsy and some non-epilepsy features and we really rarely ever see it, but it was early described, so that's fine. And there are some here that are more common and there's some that are less common, but that's what we knew. Then if you fast forward to 2014, we had a sprinkling of other genes, some recessive, some dominant. And if we fast forward again to 2016, wow, we had all these genes. So what suddenly happened is we went from, Anne and her colleagues and her neurogenetics clinic, the genetic counselors and physicians can keep track of these. We can know the literature. We can read every paper about every single one of these genes and offer expertise clinically pretty much on the fly because we knew everything that there had been published about some of these genes and we could look up the other ones so all of a sudden we're well beyond the capacity of anybody's working knowledge. All of a sudden you need not just a team at one institution, but you need a team across the world. Now I borrowed this slide from Erin Heinzen who was at Columbia University at the time and she's now at UNC and we're still working together on gene discovery. We were sort of told a few years ago, gene discovery is gonna be over, so work on functional characterization of animal models and so on. Both are quite worthy pursuits, but gene discovery still continues and we sort of went from 2016 onward to 2023 and I stopped trying to make a slide because it's just there are too many and there's over a couple of thousand citations if you look at PubMed for epilepsy genes. There's new genes coming out. There's new syndromes. There are new phenotypes associated with known genes which is what we as clinicians can contribute to, but I will ask maybe what some of you might be thinking is how can we make sense of all these data together, particularly those of us who are practicing epileptologists who are not gonna know 2,000 genes like the back of our hands. So how do we make some sense of it? One way that I like to make sense of it and this is really based on sort of my operational experience and coming from the era of 20 or so years ago where we finally had MRIs, we had patients coming with epilepsy. They had normal MRIs or they had non-normal MRIs, right? And if they had non-acquired epilepsy and they had an abnormal MRI, most of the time they had a brain malformation. So in my mind, there were neurogenetic malformation associated epilepsies and neurogenetic non-malformation associated epilepsy. And I say neurogenetics a little bit glibly, we're assuming the non-acquired are genetic in origin or have genetic contributors. And these were the categories and we said, well, here's the, you know, there's the malformations like the one I showed you with hemimegalencephaly which might have genes like AKT3 or deptic five or any number of genes with variants present in the tissue if you could be so fortunate as to get the tissue and send it for sequencing. You might have some familial syndromes like familial polymicrogyria where you have some genes involved. There were patterns seen of bilateral, frontal parietal polymicrogyria, all kinds of patterns where people said, oh, that's a patterning gene, there's gotta be a patterning gene involved and genes like GPR56 and others were discovered based on those clinically ascertained patterns in the MRI. So great, and then we had all our non-malformation genes, the sodium channels, SCN1A, SCN2A, the potassium channels, KCNQ2, some of the ones I showed you we found in the babies and you also see in other settings. So fantastic, we've got this list and we've got that list. Panels were developed by several companies except this happened. It's not so simple, actually. It turns out that there's a growing list of genes where you can have variants in those genes give you malformation-associated epilepsy or non-malformation-associated epilepsy. Deptic five is one of them. We saw it in families with focal epilepsy. There's a number of different syndromes of focal epilepsy where you have the family history is maybe a grandfather, maybe an aunt or uncle, maybe some cousins and everybody had a normal MRI and I would pause there and say, well, make sure you know that those MRIs were normal. If those MRIs were for 15, 20 years ago, you probably should look at them again. They may have some subtle findings, some subtle dysplasias, for example, that would have been missed by earlier MRI or just weren't known to be present. And then we also have genes, GRIN2B, one of the glutamate receptor genes. How does that give you a malformation? I can't tell you how, but it does and it's been very clearly associated now. Likewise with SCN3A and one of my favorites, SLC35A2, which in the germline case, it's an X-linked gene. Girls can have a severe early-onset epilepsy in the first year of life. Only seen in girls, it's thought to be lethal in males. If it's in the germline, meaning present in all the cells, you can find it in a blood test or a cheek swab test. However, if you look at small malformations and small focal cortical dysplasias, you can find somatic mutations present in some subset of the cells in those. So this categorization of, wow, these are two separate things, we can make sense of them, made sense until it didn't anymore, but still not an unreasonable way to start your sort of thought process, but keep in mind that the lines are blurring and you may wanna take a more broad-based test. And so I mentioned this here, one of the learning objectives was to think about what type of tests do you wanna do? When all we had was panels, we could do a brain malformation panel or we could do a panel for epilepsy genes that was somewhat limited. Now we can do better than that. We can actually do an exome or a trio exome to get the parent's DNA there and analyzed and not necessarily have to have such a fixed idea that this is gonna be on this list or that list. We still have to know our phenotype. And I'm gonna keep saying this over and over again. You still have to know what did your patient have? What was their family history? Does that pattern make sense with the known pattern of that gene? That's key in interpretation, which is not gonna be something that the testing lab is gonna be able to do 100%. It's gonna rely on us as clinicians to complete that picture. So what are some other categories? So we can go by syndrome. I'm using an old pre-ILAE revision term here, but it's one that many of us know, which is that it's an early infantile epileptic encephalopathy starting in the first few weeks or months of life, usually weeks. There's also epilepsy of infancy with migrating focal seizures, also very early onset. And then we've got our more common of the rare infantile spasms that I mentioned before. And we also have Dravet syndrome. And I'm just gonna point this out here that if you look at all these genes for DE, and this is now a summary that was put together by Amy McTague and Ingrid Schaeffer, Cat Howell and others, you can sort of get a sense that certain syndromes are associated with certain genes. The genes that are listed in these boxes in black are the ones where there's the most certainty, the most strong genotype-phenotype correlation. But you'll see that things like the infantile spasms, for example, there's not a strong association with one gene. There's a whole long and growing list of genes. If we just zoom in on a couple of these syndromes now. If we think about Dravet syndrome, I think most of you would know, but in case there's students in the audience, I'll just reinforce. There's Dravet syndrome, also called severe myoclonic epilepsy of infancy, presents in the first year of life, usually around six months to 12 months of age with hemiclonic seizures that occur in the presence of fever in most cases, progress to involve generalized hemiclonic seizures, some with fever, some without fever. And well over 90% of the time, if you do sequencing, you'll find a de novo variant, meaning a non-inherited variant in the gene SCN1A. But half of them are non-sense variants, meaning they cause a premature truncation. The other half are missense variant that disrupts the sodium channel function. And this is something we've known for a long time. It actually was known a little over 20 years ago, it was reported that you can have these de novo, non-inherited changes in these babies. But there were also children like that seen in families where some people just had febrile seizures. Seizures with fevers, short febrile seizures, between the age of six months and six years, garden variety, simple febrile seizures, you wouldn't work them up in any other way. But other individuals in that same family might have seizure with fever, but lasting until age nine or 10. That's not simple febrile seizures anymore, that's something else, that's febrile seizures plus. Or maybe they had febrile seizures, but then went on to have afebrile seizures as well. And you'd have to see these pedigrees with generalized epilepsy with febrile seizures plus, which renamed as genetic epilepsy with febrile seizures plus because, A, we had evidence of genetics from SCN1A, B, they're not all generalized. They can be focal, and so we had to rename our syndrome. That's sort of a long explanation that's not specific to DREV-A, but SCN1A. And I mention that because, again, DREV-A syndrome, pretty rare, one in 17,000 or so individuals. But this SCN1A really extends to a number of different conditions where febrile seizures are involved. So fine, so that's our short list. If you have a child with that sort of presentation, you wanna look for SCN1A. In the old days, we looked at single genes, then we looked at panels, now we can look at exome and trials and get the parental data and right away know, not right away, but in a few weeks know, did they have a change? Is it predicted to be abnormal? Is it from one of the parents or not? And have some evidence to put together with a clinical phenotype. Now, what if we don't find a change in SCN1A? What else could it be? On a girl, it could be PCDH19. You can get clusters of seizures with fever with that. There's this other list of genes that you might see involved with a Dravet-like presentation. Usually, if you dig into the history as a child neurologist and epileptologist would do, you'll find some differences. It's not usually a classic Dravet phenotype, but it's worth thinking about. And again, if you're getting a referral from a pediatric colleague who might not be phenotyping with the ILAE criteria in front of them, it's probably worth thinking very broadly about doing this sort of sequencing. You could do with a panel, but you may also want to do with an exome if it's faster or more readily accessible to your patients. Why do you want to know? I already mentioned the sort of, people want to know because it's medically important. Scientifically, of course, it's interesting, but you can also do something about this, right? You can generally avoid sodium channel blockers like limotrigine and trilepto or carbamazepine, which you wouldn't want to use because it can worsen it. And now you might want to know because a patient could be eligible to be in a natural history trial, that's nice, but now there's actually a gene-based therapy trial looking at antisense oligonucleotides, a very, very specific treatment that's targeting molecular machinery that can increase the amount of normal SCN1A that you would produce. So what we typically think of with Dravet syndrome and SCN1A is that it's a haploinsufficiency. You're missing half as much as you ought to have. So if you can actually inject into the spinal fluid, this sounds a little scary, it is a little scary, but you can inject into the spinal fluid a molecular compound that's going to go find with its little molecular GPS, find the neurons and find the SCN1A area and force the cells, force those neurons to make more. You might fix the molecular defect that's been done in cells. It's been done in mice by Lori Isom's group in Michigan and others. And it's now in a clinical trial. This isn't easy. This isn't a walk in the park. If you have a patient with Dravet syndrome who is well-controlled on valproic acid and the ketogenic diet and clobizam, stop. Maybe they're well-controlled on fluramine and epidiolex cannabidiol, stop. You don't need to go to a high risk therapy, but if your child, the patient is having tons and tons of seizures and nothing's working, there is a clinical trial. And if you know the gene, you can be eligible for the trial. If you don't know the gene, you're not going to be eligible for the trial. And so this is one of these examples where it's, you know, just really important to know. So that's Dravet syndrome. What if you have infantile spasms? Are you going to go for a single gene testing? Absolutely not. And are you going to go for a panel? Well, you can, if that's all you can do. But the panels just, every time a new one comes out, it's going to be behind because we keep learning about more genes, as I mentioned. So I think the best approach is a broad approach. If you can do it, I understand some practice settings don't allow you to do it. So maybe a free panel with just the child is good enough, but really and truly, if you can lobby for exome with trio data, whenever possible, you'll get a much more comprehensive picture. You'll get a much more comprehensive answer. I said, I wouldn't say a whole lot about this, but I just want to mention that this, because it is common, we do see focal cortical dysplasias, for example, all the time in any epilepsy practice. I mentioned the hemimegalencephalies. We've actually now, you know, between 2012, when we first started reporting some genes for this to now, we've actually been able to solve about 80% of these cases. If you look at the brain tissue with mTOR pathway genes. So that's important if you've got patients having resection and they've got a huge lesion like this, you take a piece of the tissue, you can do sequencing. There's a couple of commercial labs, but only a couple in the US that actually do the sequencing. You can send some cheek swab DNA and send some brain tissue and they'll take the DNA out of the brain tissue and they'll find it, which means Alyssa and I and our colleagues don't have to spend our research time doing that. We can all start thinking about how do we treat these kids better? And maybe how do we make these diagnoses without having hemistherectomy happen, right? I mean, it's a little bit too late though. And so we're looking at ways to look at maybe cell-free DNA and the CSF, which many colleagues around the world are trying to optimize. Maybe if these kids are having recording with subdural electrodes, can we get some DNA from those and make it a molecular diagnosis? The answer is theoretically yes. The practicality is that even if you know what you're looking for, it's pretty hard. And so we have some optimizing to do, but what a world of difference from what we used to tell families, which is we don't know, now we know, and now we can pursue it. And because it's mTOR pathway genes, we can learn from our colleagues who've been treating TSC for all this time and look at some of the genes that they've, the gene-based treatments they've been looking at and try to adopt those in these kids, perhaps before surgery. So I'm hoping we'll have clinical trials soon that will target these types of kids with lesions that are possibly mTOR inhibitor amenable. So that's that group. Focal cortical dysplasia, we went from almost no causes. There were some rare syndromic cases. There's an Amish founder allele in the gene CNTNAP2. Maybe you've heard of it. Maybe unless you live in Lancaster, Pennsylvania, it doesn't matter. But for the rest of the world, it doesn't really show up as a genetic defect, except when you look at the tissue. And then we found some mTOR pathway abnormalities, particularly in cases that look like this, like a type two dysplasia with this wedge shape, and it goes all the way down to the ventricle. But in the FCD type ones, these little blurry lesions that you almost miss the first or second or third time you look at the MRI, but wow, the EEG keeps pointing to that thing. Locus, let me look a little harder. Do some additional scans and you find it. When that tissue comes out, it's when our SLC35A2 gene will show up more often than not, particularly when the pathology shows a clonal expansion of oligodendrocytes, which is non-tumor, but somehow related to this gene, which is involved in galactose transport. So all sorts of new biology. Never would I have dreamed that a malformation like this would be caused by a galactose transporter. So it just, it sort of shows you that, whether it's a genetic defect, it sort of shows you that whether it's in the research world or in the clinical world, we have to kind of keep eyes open. And that's sort of why I continue to argue for a broad approach in terms of look at as many genes as you can. Other genes keep showing up in the brain tissue, PCDH19. Anything that can be mutated will at some point, and it might show up in the form of focal epilepsy. We just haven't solved it all yet. And then we've got a number of cases where you've got essentially normal cortex, but in temporal lobe epilepsy, we've seen BRAF, KRAS, cancer genes, mTORopathies are cancer genes. My arrow's not working here, sorry. The mTORopathies are cancer-related. These TLE genes are cancer-related. Our friend and colleague, Dr. Kushko, has been working on this in Chris Walsh's lab for a long time and found some really cool and really surprising findings that good old temporal lobe epilepsy will often start in childhood, right? They'll have febrile seizures. They're going on into adulthood with temporal lobe epilepsy, which is the most common focal epilepsy in adults. What happened between that focal seizure to adulthood? We're not really sure, but something must have been going on in that brain region, in that vulnerable temporal lobe. I thought 20 years ago, well, most of the cells that would have had anything abnormal with them would be gone, right? Because what happens is musculoskeletal sclerosis, you'll lose a lot of tissue. But there's some left that have been ascertained in assay, and they have these oncogene variants. So stay tuned because I think this is gonna be a huge game changer in temporal lobe epilepsy that's sporadic, it's not familial, and it's caused by variants in genes that we know all about from the cancer world, but that we never really consider to be relevant to us. So again, to keep a broad view, you'll end up looking and finding a number of different things. All right, so we've talked about some genes by name. We've talked about some other genes in broad categories. I've shown you what you can see in older kids and in babies. It's time for an audience participation question, and we have a nice little poll that should pop up. And I just want you to look at the pairs here and to see which of the following pairs reflect an established gene disease association, where we know enough about the gene and we know enough about what the phenotype can look like, that if you found a patient with a variant of one of these genes and they had this phenotype, you'd say, yep, this could be it. Now, I can't see the whole poll. I can only see the first, oh, there we go. All right, I can see so far we've got, for the A, we've got 91% of the people saying yes for the B, a handful, C, a handful, and D, most. Now the poll is closed. All right, so let's go through this. Dravet syndrome, SCN1A, if I didn't beat anything else over everybody's heads, hopefully that was it. So that's a strong association, so great, excellent. Rett syndrome and DEPTIC5. Rett syndrome I didn't actually talk about so much, but Rett syndrome typically is associated with MECP2, and it's typically associated with a loss of function in MECP2, although there can be a duplication syndrome also. Focal cortical dysplasia is more often associated with DEPTIC5, which is upstream of mTOR. However, if you said focal cortical dysplasia in MECP2, I did just say like 30 seconds ago that if you look broadly, you'll probably find something. So I would give you that as kind of a half point. It's not the typical association, but probably somewhere out there, somebody with FCD has a variant in MECP2, and it could be causing a local disruption. So I've learned something from this poll, which is that I didn't really design CST very well. And then also that the neonatal onset of left-foot case CMQ2, yes, absolutely. That's one of those genes that showed up in our first week of life, but also in our first month or so of life in our genome example with 100 kids that were sequenced, and this is sort of a prototypical early onset, neonatal onset seizure gene. It can be tonic seizures, Odahara syndrome, birth suppression, EEG, really, really terrible prognosis. It can also be a self-limited neonatal or infantile focal epilepsy. Take a family history, ask the grandparents, did your child ever do this? Sometimes the grandparent will be the one to say, oh, yeah, this looks familiar, but it went away. Were we doing exome or panel testing when the parents were little, 20, 30 years ago? No, we didn't have those tools in the clinic. And so the history is sort of everything in terms of how the prognosis could play out. All right, moving on. I wanted to talk to you a little bit about how we've tried. You know, I run a research lab as well as a clinic, and I mentioned I'm only in clinic once a week. And like probably many of you who are seeing patients and working hard in the clinic all week long, we're really trying to sort of discover genes, but also figure out what they mean and how to interpret the findings in those genes. And many years ago, Brandy Fuhrman, who was then at the NINDS and is now at the Epilepsy Foundation, suggested that we kind of write up how we do this. We kind of go back and forth from the clinic to the lab and take the discoveries from the lab over to the clinic. But more importantly, we were taking our observations from the clinics and the lab and saying, well, this family's epilepsy looks like this other family's epilepsy. Maybe we should look for the same genes and putting pieces together with colleagues from all over the world. So we had this translational model that we put together. And really where that led us is to some of the findings that I showed you and the observations in all of those kids, the 20% of the epilepsy all comers, over 40% of the babies with epilepsy. And what Beth Shively and our group set out to do was to say, well, that's fine for our group, but let's look across the whole field. It's not just that there's this handful of genes. Now there's all of these different studies from all over the world looking at populations of kids, sometimes hospital ascertained, sometimes population-based, which is really the best way to study in terms of epidemiology, but what were people finding? So we undertook a systematic review, which really showed us that in fact, there is actually good yield from testing. On the order of 20 to 40% of exome positivity, higher in kids who had younger age of onset, higher in adults who had pediatric onset that was younger in age, and highest really in the population of individuals with intellectual disability. So I'll preempt a question that my colleagues always ask here. It's like, do you mean that the kid with absence epilepsy who's well-controlled and at the sex of mine, we should be doing trio exome sequencing? And from a resource utilization perspective, and really does anybody need to know perspective right now? Probably not. You can tell that family that there's a 5% risk that a first degree relative would have epilepsy in a case like that. So for that child's siblings, that child's children eventually, it's about 5%. By the time that child who's probably five or six years old, if they have absence epilepsy, by the time they're an adult and hoping to have kids, I really hope we understand the polygenic risk that is behind absence and some of the other generalized epilepsies. But are you gonna spend time getting a prior auth and doing all those things? Probably not. But if they're not that kind of textbook, classic garden variety, otherwise healthy kid, it makes sense to do it. And there's also non-yield outcomes, personal utility, utility for the family, and those things which were not well studied. So we couldn't actually comment on those, but we could say, we could make recommendations about testing, which we as physicians didn't do, but our genetic counseling colleagues did. So Lacy Smith, who's here at our institution, worked with Beth and others. And the National Society of Genetic Counselors said, you know what? The professional neurology associations are fabulous. And they have these wonderful ways of putting together guidelines based on systematic evidence review data. But they take a long time. We just wanna do this. We want our patients to get this testing. So they put together guidelines, and then, sorry, there's your yield, which I mentioned already. But they put together guidelines, which we then were able to adopt and endorse through the American Epilepsy Society, which I'm very pleased about. So what do we do in clinic? As I've mentioned, I think a few times, I try to take a general approach. You have somebody with unexplained epilepsy, confirm it's epilepsy. We take histories, that's what we do, right? Make sure it's really epilepsy, get the data. If it's really unexplained, then take a general approach. Most of the time with exome genome, if you can, we can't clinically. And do a trio if the biological parents are available. Now, if they have dysmorphic features, by all means, look for copy number variants, right? Deletion syndromes, duplication syndromes with a chromosome microarray analysis. And you may do panels if that's what you have available and nothing else is available. Now, you can use these tests in a complimentary sort of stage-wise fashion. Let's say you do an exome, because that's the easiest thing to do. And you don't find something. Is that gonna tell you if you have a deletion of chromosome 1P36? Actually, today it should. A few years ago, it didn't. But now, actually, the labs can interpret fairly large chromosome deletions and duplications based on the exome data. So we're in luck there, because we're kind of getting a two-for-one. But let's say you see a patient, and you say, well, this child is a girl with microcephaly, and they're not developing well. They have focal seizures, and now they have infantile spasms. You know, boy, I think this might be CDKL5 deficiency syndrome, but my exome was negative. Maybe you do wanna do more dedicated testing, or maybe you wanna do deep sequencing and sort of pursue other things. But maybe they don't have such specific features, and you sort of say, oh, I do my exome. I don't know what else to do. I'm just gonna do a microarray, because there's an 8% yield there in epilepsy, so let me take a look. Or I'm gonna keep reanalyzing that exome every year. So you can take a mixed approach as well. If you have a specific hypothesis like Gervais syndrome, you know, continue to address that hypothesis. If the exome is negative, maybe do a deep panel. Maybe they have a mosaic SCN1A variant present in only some of the cells that the usual testing won't find. Or call the clinical lab and say, hey, you know, I really think this is SCN1A. Could you look, could you see if that gene was covered well? Did something happen? And just make sure it really was saturated, that every ATCG base pair got covered there. And then finally, I'll end by saying I really think we are looking at our patients one at a time in clinic, but we really can take a collective approach. So if they have focal epilepsy, let's say, we can solve what they have, either with blood, cheek swab, maybe brain tissue. And maybe we can't solve what they have, maybe we'll be able to do expression profiling. I didn't talk about this. We don't do this clinically, necessarily, but maybe one day we'll be able to take a tissue and look at not just what genes are there in their letter code, but what's being expressed, what genes are expressed. And we'll find different patterns. We might find mTORopathies and treat them with mTOR inhibitors. We might find galactose deficiencies or transporter defects and treat those. We might find rasopathies and TLE and treat those. And we can maybe find those and treat them in similar fashion. But I'll warn us all, including myself, that it's not just gonna be three colors. There's lots of different possibilities. And so I hope I've given you a little bit of a sense of kind of where things are and where they're going, and hopefully some excitement that we should try to do this. So we'll go on now to the next poll question. This is a case scenario of a seven-month-old girl, history of focal seizures from age four months, now presenting with infantile spasms. Which of the following would be most appropriate first-line test to achieve a genetic diagnosis efficiently? Would you do single-gene testing for MECP2? Would you think about epilepsy panel testing for the girl? Would you think about chromosome microarray analysis or maybe triaxome for the girl and her parents? And you can pick whichever one, you know, suits you based on your practice setting. You might be able to pick more than one, I'm not sure. And let's see, I'm gonna... Can I broadcast the results of that? Okay. I think they're all mostly in. Nope, where did they go? Can you see them, Dr. Moldofer? For me, they're at the bottom right of the screen. I hope our participants can see them. I'll just show you. The prevailing result was, actually almost a tie, but 46% epilepsy panel testing for the girl. And almost as many, 42% said trio sequencing for the girl and her parents. Both of those should give you a, I would think, I was leading us towards a CDKL5 X-linked pathogenic variant that would be the novo in the girl. You know, if I said microcephaly and all these things, it's not impossible that this could be MECP2, although it's not typical that they have infantile spasms. And so you probably wouldn't wanna spend weeks on single gene testing there. And that's why that's not the answer. But either B or D ought to get you the answer. The reason I still prefer the exome, if you can get it approved and you have a way of doing that is because if it's not CDKL5 and it's something else, and it's not something that's on your panel, you don't have to then go to the next step and do something new. You can just go back to the data, right? And the analysis will do that. And now if those are negative, you may think about chromosome microarray analysis. What if it's a small deletion in the gene deptic five, which is associated with infantile spasms or a small deletion in dynamin one, which is associated with infantile spasms or any one of those other dozens of genes, very possible. The chromosome microarray analysis will get you that with more resolution than an exome. So you might think about that. But depending on the practice setting, you'd probably either do the panel or a trio exome. So takeaways, I hope that you're convinced now that non-acquired epilepsy should be considered genetic until proven otherwise. We've had a multitude of studies and accruing evidence and now systematically reviewed evidence from the last 10 or 15 years. Evaluation should be based on hypotheses from the clinical presentation. You think hypothesis, you're thinking, oh, are we talking about a grant proposal? No, this is clinical hypotheses. We do it all the time with every diagnostic adventure we take, right? This is what I see. Here are all the clues. Here's what I think this is. Now let me pursue it with some data. Options include testing for single nucleotide variants as ATCG letter changes that can give you non-sense or missense variants, but also for copy number variants. Unsolved today doesn't mean unsolvable. So if you find something or you don't find something and you don't have an answer, but you really think, wow, this patient was a young child who showed up with regression and epilepsy. I hypothesize that this is genetic. Keep looking. It doesn't mean that it's not solvable. Maybe you just reanalyze your exome data. Maybe you bring genome to the picture when you can. The yield for younger patients and patients with intellectual disabilities approaching 50%. It's not 100% and people will sometimes push back and say, well, you're not gonna diagnose all of them. When I started, we were diagnosing none of them because we had nothing at our disposal. We've made a rapid increase in our capacity and we're now trying to show not only the yield, but the long-term outcomes. But keep in mind that the field is changing and that genetic results really do change management and prognosis when the hypotheses are well outlined by all of us and the data are robust and well-analyzed. So with that, I'll just say thank you to our participants today. Thanks to all of you and our patients and families, especially local and other funding sources. And then to our lovely team, some of whom I think are here on the call. And again, to our patients. We certainly enjoy what we do together, but it is absolutely a team sport. So with that, I'll welcome any questions. And just as a quick reminder, text chat is located on the right-hand side of your screen. To submit a question, type your question in the small text box at the bottom and when finished, click the send button. And now I'll turn it over to Dr. Milhoffer to facilitate. Yeah, thank you again for a great, great overview. It's such a complex topic, but amazing to see how much has changed over the years. I wanted to get right away to the questions that came through in the chat here. The first question here from Charles Neeson was, can you give an example where a novel gene finding in an epilepsy disorder led to a gene-specific improvement treatment? Yeah, I gave one metabolic example, regression, generalized seizures, and that's the CAD or uridine deficiency syndrome. There are a couple of others. There's certainly a pyridoxine dependency, which maybe you could make at the bedside by just giving pyridoxine. And I would encourage everyone to still do that if you've got particularly unexplained neonatal epilepsy or infantile spasms. But other times it might be an atypical presentation and you'll come across a B6 deficiency or something of that nature. Most of them still, the ones that are quickly, readily treated like that are going to be these deficiency syndromes. But there are others in a growing list where you can have a specific disorder such as leuphora disease or other progressive myoclonic epilepsies, Batten disease, and so on, where there are now gene-based therapies kind of going to trial. So work from Berj Manassian has been really exciting with leuphora disease. And they went through many, many years of natural history and many years of development, but now have some AAV viral vector-based therapies. There are antisense oligonucleotide therapies developing. Those are not available to everybody. And I think we have to be honest about that. And I think they will be if they're effective, but we still have to show that they're effective for seizures and development. We have to de-risk them a little bit. And I feel like our job when we're in the clinic is to let's get those patients diagnosed and lined up for when they're ready. So it's still a minority who are having a game-changing precision therapy, for sure, but I think it's a growing number. And so if we keep the patients ready, then as the treatments become available, we'll have them. Oh, okay, great. Thank you so much. There was a more kind of practical question for someone that is in private practice, where to order genetic testing and which panels should be ordered. I think you spoke to that it's somewhat driven by the prototype as well, but is there a specific company? I'm not quite sure whether you can promote anything here, but how you would approach this in private practice? Yeah, it is very practice-dependent, right? Certain institutions have contracts with some testing companies and so on. So it is very practice-dependent. I was talking to colleagues at a practice in California recently and they said, oh, all we can do is this, but actually we can take the data that we're learning and try to advise the practice. So there are free panels available and many institutions will not allow you to do send out exome sequencing because they lose money on it because it doesn't get reimbursed. So I think if all you can do is a panel, a free panel, which is just the child, you can start there. You'll at least get, I mean, to Dr. Neeson's question, you'll at least get the treatable ones covered, right? You'll get a CN1A where you're gonna direct treatment away from sodium channel blockers and maybe a trial. You'll get the B6, you'll get the GLUT1 deficiencies hopefully there and can try to treat with a ketogenic diet. So you'll get sort of the low-hanging fruit of the treatable. Many of those cases, if it's a variant that's not a known pathogenic variant though, you're gonna be stuck then getting the parent's DNA and making sure it's de novo and probably working with a geneticist or a neurology colleague who does neurogenetics to prove that that's really the cause. So there's a little delay in a secure diagnosis when it's not a known previously described variant, but you can start there. Or you could take the data from conferences like this and presentations like this and sort of try to get your institutions to look at the data. The data are really compelling and you can actually look, I didn't present this, but there are a number of cost-effectiveness papers too saying if you start with a bunch of 100 kids and you start with a trio exome, yeah, you're paying for 100 trio exomes, but then you're done with almost, if it's babies, almost half of them, right? And the others, you're gonna do reanalysis of that and maybe you'll do other testing. If you start with a panel and you're only diagnosing, let's say, 15 of them and you then know that there's gonna be at least another 30 or so, you're then gonna do exome on all of them again. So it's just from a cost perspective. If we actually run the numbers, it's cost-saving to actually do the trio exomes today. Not genome, it's super expensive, but those costs will change. And so some of the literature, there's a paper from Catherine Howell in Melbourne, Australia who's put it into their system, but it's applicable to ours. And then a paper from Yvonne Sanchez-Fernandez in our system looking at, and actually with a little app where you can plug in your own yields and plug in your own costs from your own population. But I think it's on us, for those of us doing this work to sort of continue to demonstrate the cost-effectiveness. Okay, great, thank you so much. There was a specific question about whether, how do you kind of approach or determine whether we're dealing with a gain or loss of function, for example, in GRIN genes? This question was raised by Dr. Kharazian. That's a great question. And it's really important because some of the gene-based therapies are actually really quite dependent on that. The GRIN ones are a great example. If you have a GRIN2A gain of function, for example, you might be able to suppress that with an available compound called memantine, which has been used in Alzheimer's, but it's also been used unsuccessfully for things like ADHD and things in kids. It's safe in children. So you might be tempted to try that, but if you have a loss of function, you really don't wanna make things worse. And so how do you know? In many cases, you really do need functional data or an annotation of functional data. So the example, that's my favorite example, actually, because Steve Trinales at Emory has put together a database, the CFERV database, Center for Evaluation of Rare Variants, where his lab has actually gotten funded and characterized one by one. We're talking about hundreds of variants, but they've characterized many of them according to gain of function, loss of function properties. So when I see a patient in clinic with a GRIN2B or a GRIN2A or so on, and I don't know the variant, which most of the time, because I can't keep all that in my head, I'll go to that database and I'll look and I'll say, oh, wow, this is what's happening in the presence of glutamate. This is what's happening to glycine. And if I don't remember exactly what that means, I go to the one I know is gain of function and I'll compare and contrast and I'll say, this looks like one where we should try mementine or dextromethorphan and so on, versus, wow, this is the opposite direction of those published cases. Let's steer away from that. And I'll be honest, I'll often just send an email to the authors of the papers and say, look, I'm seeing this patient. I really wanna try this. It looks like there's an orally available therapy, but I don't wanna take a chance that I'm gonna put things in the wrong direction. So I'm pretty conservative in that regard because we don't wanna do harm. I wanna use the information that's available. It's an argument for making more of these data publicly available. Yeah, that makes absolute sense. There was actually an interesting question about specifically about SCN1A. And I assume we can extend that to other genetic mutations that you mentioned. Can they be also found in people with normal phenotype and without it, is that a possibility? It is possible. It sort of depends on the gene and the type of the gene and the variant itself and where the variant is and how it's affecting the protein or not. And then the truth is that there are some genes where that observation exists. SCN1A is one of them, like those GIFS plus families. There are skipping generations where people have the variants, but they may not be affected. Deptic 5 is another one, SCN3A is another one. Whereas there are other genes where you just simply never observed that. It does make the interpretation tricky. And that's when I talked about that needing to be careful in terms of the variant interpretation and bringing our clinical lens to it is also have to sort of bring the collective clinical lens to it. If you're seeing a variant in a gene that has been described as highly penetrant, meaning anybody who has a change in that gene that's not a population variant, has a phenotype, be hard pressed to say that, oh yeah, that person's child might have something that's gonna be the explanation versus some of these other genes where we know it's happened, then we can use those data. So it's really this pooling of all these different data sources that goes into the interpretation. That's interesting. Also, Dr. Berg had a question about the observation that mutations in multiple genes can result in similar or same phenotypes suggesting that these genes might have a functional relationship. And he was wondering in channelopathies such as Dorvay, do the known disease causing genes co-localize or are they known to be in the same network or have some other relationship? And are there some patients with multiple partial loss of function mutations that result in severe phenotypes? And how does that influence the use of genetic testing? And that's the context. I'll do the second one first, which is that we do think that there's a polygenic contribution to some of the milder epilepsies that we haven't been solved. In general, if you have a severe epilepsy, you have intellectual disability, it's a severe phenotype and it's gonna be so severe it's gonna impact your ability to reproduce. By and large, what we've seen is variants, single variants in one gene that are causing a huge effect on the function of the resulting protein. So one gene, one condition, whereas the milder conditions are where we hypothesize that there's many small changes in many genes. Now you could have, let's say, a sodium channel variant, SCN1A variant that's pathogenic and a bunch of other variants in other genes that might drive you towards a more severe phenotype. Maybe it'll be Dravet-like, but even earlier onset. So I can't say that there's not an influence of other genes, but usually there's one that's the main driver. We're not going and looking for other ones because we don't seem to need it. It seems like it's sufficient, and it's sufficient in animal models to produce seizures and cases like that. So we're not sort of needing to go look. Whereas with the milder epilepsies, it's probably a polygenic risk. And we finally have data now. There's some really nice papers from large consortia where we have data to put together what's called a polygenic risk score and say, okay, this person has X number of variants and X number of genes that we know are epilepsy genes. They have a high risk of epilepsy. That hasn't yet translated to single patient, take a sample, let's see what they have and score them. But I think we're getting quickly into that direction. The other question was about the many different genes consolidating to one phenotype or sort of coalescing into one phenotype like hypsarrhythmia and infantile spasms, for example. That one's every different. You can have a synaptic disorder, an ion channel disorder, an early developmental disorder. It's really in almost any pathway. Dravet, it's a smaller list. It tends to be sodium channel GABA receptor subunits, but good old PCDH19 is affecting brain development. They're all affecting circuits. SCN1A we know is really important in interneurons and it's interneuron-specific. Interneuron-restricted mouse models can have the same phenotype. And so are these other disorders all interneuronopathies also? I would suspect there's some effect, but we don't know that yet. Great. I think we're coming to an end here. We're a little bit over time here already. But there was one question. You recommended reanalysis as initial analysis might not resolve the question. So is there a specific time frame that you would recommend to wait before you reanalyze? Yeah, the reanalysis, yeah. You're sort of allowing new knowledge to influence the analysis of a variant and or maybe you can get parental data if you didn't have it originally and so on. Many of the clinical labs will offer one free reanalysis of an exome or trio exome after at least a year. So I think you sort of want to wait at least that long. Or you can send it to colleagues who do research reanalysis and can sort of do it anytime. But from a practical perspective, about a year or so later is enough. If the child is not doing well, they're sick and they're in the hospital, call the clinical lab and say, hey, this child's not doing well. Can you take a look again right now? And they will. Most of the companies are very responsive. Okay, well, that's great to know. All right, well, I think I wanted to say again, thank you on behalf of AS and trio speaker and all the participants for today's event. A recording of the webinar will be sent to you within the seven to 10 days. And please make sure that you complete your evaluation in order to claim CME for today's event. And that will conclude today's webinar. And thank you again, everyone. Thanks everyone for coming. And thank you, Dr. Mulhoffer for moderating.

genetic testing

epilepsy

Dr. Anne Paduri

neurology

genetic sequencing

treatment decisions

clinical trials

anti-sense oligonucleotide therapies

individualized patient care

standardized approach