H3 K27M–Mutated Diffuse Midline Glioma: Bridging Molecular Pathology and Clinical Care

This transcript was automatically generated from the audio recording and may contain inaccuracies, including errors or typographical mistakes.

I am Michael Ruff. I am a neuro-oncologist at Mayo Clinic. I primarily see adult patients. I certainly see these patients with the H3 K27 alterations. Dr. Horbinski, I have collaborated with you indirectly in a few instances, but it is nice to have a discussion with you.

Dr. Craig Horbinski (Mayo Clinic): Very good. Looking forward to it.

Dr. Ruff: We want to discuss how H3 K27M alterations have transformed the diagnosis of diffuse glioma. This has been taking place during my training. I think right when I started fellowship was when this entity had come to light. It continues to evolve, and much of it I am familiar with; the entirety of it, I think, you are probably much more familiar with.

Dr. Horbinski: Yes, the notion of the K27M. When I was a trainee, we were still making the diagnoses exclusively by histopathology. Back then, we called it diffuse intrinsic pontine glioma because, clinically, radiologically, and pathologically, we recognized that a subset of gliomas, thankfully a very small subset, arise in the brainstem, whereas most gliomas are going to arise in the supratentorial in the cerebrum. However, these would arise in the brainstem, mainly the pons, and that is why it was called diffuse intrinsic pontine glioma. We knew, of course, that those were terrible tumors, extremely poor prognosis; the only thing you could do with them was radiate. Median survival was roughly six months or so.

Some time ago, some big genomic studies came out showing that these tumors have this odd mutation in a histone-encoding gene called H3F3A. It was always K27M. That became the diagnostic criterion by which you called something a DIPG. Then, as time further advanced, we discovered this is not just in the brainstem; any glioma arising in the midline, even up in the supratentorial space, could harbor this mutation. Eventually, then we said, "It is silly to call something a diffuse intrinsic pontine glioma when it is arising north of the pons. Let us change the term to diffuse midline glioma or DMG." In some of the older literature, you may still see DIPG used as an abbreviation, as an acronym. It is the same fundamental disease process.

Dr. Ruff: My understanding is, it is loss of trimethylation; so the mutation causes this diffuse loss of methylation throughout the cells that are not necessarily mutated. Do I have that correct?

Dr. Horbinski: There is an emerging subclass of mutations. IDH is another one where it is in the old days we had tumor suppressors, and we had oncogenes tumor promoters, meaning if a mutation occurred in, say, p53, well, that is a tumor suppressor gene. You are inactivating p53. Ergo, you get a tumor. You get an activation or amplification in some fashion of, say, EGFR. That is an oncogene. You are putting your foot on the gas instead of taking your foot off the brakes. That is the oncogene versus tumor suppressors.

Now, this emerging class of mutations are occurring in these genes that you cannot call them oncogenes; you cannot call them tumor suppressor genes. These are epigenomic modulating genes. Histone H3-3A is a histone-encoding gene, and the protein is H3. When that is mutated, it does not just affect one gene; it affects pretty much the entire genome. It is reshaping the epigenome of that cell in such a fashion that it is facilitating through ways we do not clearly understand. It is facilitating tumor formation. Likewise with IDH mutations. The IDH gene is not a tumor suppressor or oncogene. Instead, its effects, through the production of the D-2-hydroxyglutarate oncometabolite in the mutant enzyme, those effects are genome-wide. They are affecting and turning on and off all sorts of other genes that lead to tumor formation.

Now, the next question might be a fair question if somebody is thinking about this and does their homework: why is it that these mutations are restricted to such very specific kinds of tumors? For example, you can see p53 mutation in just about anything. But if I see a K27M-mutant, I am pretty sure it is a diffuse midline glioma without knowing a thing about the patient or having seen the microscopy. That tells you that it is restricted to a very specific time and place. Likewise for IDH mutations. If I see an IDH mutation, there is only a relative handful of tumors in which these occur on any frequent basis. Gliomas being one of them.

It tells us that there is something about the background epigenomics of the progenitor cells associated with these tissues that, in some fashion, makes them vulnerable to the modulating effects of these specific mutations in ways that we do not really understand yet, in part because it is very hard to model and study this on an experimental basis.

Dr. Ruff: Got it. DIPG is typically when I think of them, I think of - well, not to go back to the old nomenclature, but I think of little kids. My understanding is that is because the pons, there is some element of development in the pons occurring at that phase of life. That is where this progenitor cell picks up this H3 K27M alteration and then stays in this persistent immature state. Generally speaking, the same is true for IDH mutations in that they occur in a characteristic region and at a characteristic age such that later in life, we are much less likely to encounter them in a patient. I have often heard that after age 54, you do not routinely test for it. I am sure that is variable institution to institution, but just that these mutations occur in a specific anatomic compartment and in a defined age range.

Dr. Horbinski: Yes. In fact, that has been recapitulated experimentally, at least in the context of diffuse midline gliomas. There was a marvelous study published. Where they did transgenic mouse modulation. When they induced that K27M mutation in the hemispheres of mice during their various embryologic stages, they could turn on and off the mutation both spatially and temporally. When they induced it in the hemispheres of the mice, it did not matter when they turned that mutation on; nothing happened. The mice were fine. They were fine. No tumors, no nothing. In the midline, when they turned the mutation on at a very early stage of embryogenesis or later, like after birth, for example, nothing happened. No mutations, no nothing. I think it was a three-day embryologic window in the mouse's development when that mutation happened. If you turn that mutation on in that three-day window, 100% of the time you got a diffuse midline glioma. That really tells us that there is something going on here.

Now, why haven't we been able to do something similar with IDH mutations? I think it is because the effects of D-2-hydroxyglutarate are much more complex and much more delayed. In other words, the IDH mutant effects on the epigenome and the requirements that are needed for that to manifest a tumor are a lot more complicated. It is not enough just to have the IDH mutation. In fact, we know this because we never see gliomas that only have an IDH mutation. They either have ATRX and usually p53 mutations hand in hand, so you get that troika of mutations, or you get TERT mutations and 1p/19q codeletion along with the IDH mutation. The former are astrocytomas. The latter are oligodendrogliomas. That tells us that the IDH mutation is not enough. It needs help. The only way you are going to get a tumor is if these additional mutations arise. It also probably explains why the age range that we typically see these mutations, both the midline gliomas as well as these IDH-mutant tumors, they are fairly restricted in terms of the age ranges in which they arise. Although that said, the K27M mutation, for example, I have diagnosed 70-year-olds with diffuse midline glioma, K27M. It can happen. Underscoring, in my opinion, the importance of using advanced molecular diagnostics to screen routinely all gliomas and suspected gliomas for all of these markers simultaneously. The technology is there. It is cost-affordable now, and there is no reason not to because you get surprised.

Dr. Ruff: Absolutely. I am managing a gentleman now in his 40s, where there was somewhat of a diagnostic saga seeing at fantastic institutions, and even entertaining an inflammatory problem in the brainstem, for instance. Eventually, push comes to shove, we need a tissue diagnosis. We tried CSF analysis too. We looked for tumor DNA or 2HG, the metabolite of the IDH mutation in the spinal fluid. That did not render a diagnosis, so we had to get a tissue pathology, safely underwent a biopsy under the care of our expert surgeons, and indeed, this gentleman in his mid-40s has an H3 K27M-altered DIPG. Maybe not the most classic radiographic appearance for it when I think of children with DIPG, just an expanded pons, for instance. We really did not know what his diagnosis was, to the point that we felt that biopsy was the most reasonable option.

I have had different versions of the same scenario. I have had the same appearance of an IDH mutant tumor. For a time, there was a trial here using short-course radiation where patients would receive 25 Gy in the setting of a diagnosis of DIPG, and then they were observed. Then, when the tumor grew again, treated with another 25 Gy and then observed, and then when the tumor grew again, another 25 Gy. I was trying to enroll him on that trial actually because I was so certain he had this DIPG. The PI of the trial said, "No, I think we should biopsy," and indeed, it was an IDH mutant tumor. Totally different disease, very similar radiographic appearance, and a diagnosis that we could not make otherwise, and have very different clinical trajectory as well.

I have had another example in which we did not feel that it was safe to biopsy - it was down in the medulla - similarly, tried to make a diagnosis with CSF, were unable to treat it on the same trial, and has had a course that is not consistent with an H3 K27M-altered tumor, and so it makes it difficult in terms of additional lines of treatment because I do not really know if there is any silver bullet that we have; we still do not have a tissue diagnosis, so in the absence of tissue, I am often handcuffed. I have possibly beneficial tools, radiation, alkylating chemotherapy, but I am not going to use a targeted agent. That is, I think the really exciting element about "If we can prove what this tumor is, maybe I have a good targeted agent that I can prolong quality of life and duration of life." I have just been astounded at all of the improvements and breakthroughs and increased understanding of these tumors that I would speak absolutely to the limitation of MRI. We are not certain if it is inflammation or tumor. That is a huge discrepancy in terms of the disease process.

Dr. Horbinski: One of the things that we are actually working on as we speak is a collaboration between Rochester and Florida Mayo is improved diagnostics of potential brain or suspected brain tumor lesions by cerebrospinal fluid analysis. We have an advanced molecular technique that we have developed in my laboratory. We have adapted it. Others have developed it as well, but we are testing it in cohorts of cerebrospinal fluid to see if we can pick up these either mutations directly or epigenomic signatures that indicate the presence of a particular mutation; for example, the diffuse midline gliomas have a very distinctive epigenomic fingerprint. Even if you do not detect the mutation directly, you can look at that fingerprint and say, "That is got to have a K27M." Likewise, if you have an IDH mutant tumor, you do not have to necessarily prove that it has the sequence alteration in the identification. You can look at the epigenomic fingerprint of that tumor and know for sure that it is an IDH mutant tumor. Those are the things that we are trying to suss out now from cerebrospinal fluid.

Diffuse midline gliomas in particular lend themselves to that because they are almost by definition, they are very, very close to this circulating cerebrospinal fluid, the subarachnoid space, the ventricular system. We should, on a spinal tap, have a decent shot at least of picking up diagnostic cell-free DNA. That is what we are working on to reduce these diagnostic challenges because, as you rightly pointed out, tissue biopsies in the pons and the brainstem, that is no small thing; that is what we call high-priced real estate. Very, very challenging to get diagnostic tissue in such a delicate area.

Dr. Ruff: Absolutely. Even among my surgical colleagues, they will not all advise or pursue it. There are a handful who are much more comfortable than others. To some extent, it depends on who you ask as to is it feasible or not. I am always grateful when we are able to safely get a patient through that.

Dr. Horbinski, I have a question. When you are analyzing the spinal fluid and looking for the epigenetic signature, those are still tumor cells, or that is free-floating tumor DNA, or that is like tumor-associated myeloid cells; they are all transformed, right? They would all be cancer cell genetic or epigenetic material that you are looking at with this new collaboration, new project.

Dr. Horbinski: Yes. The advantage that we have is we know what the epigenomic fingerprint of these tumors are directly. The idea then is: if we know what we are looking for, can we detect that very same signature in the cerebrospinal fluid? In some sense, we do not care if it is coming from intact tumor cells or DNA that the cells sloughed off when they gave up the ghost or anything. We really do not care. We just want to know if we can see that epigenomic fingerprint and if we can differentiate it successfully from, say, the epigenomic fingerprints associated with inflammatory cells and those sorts of things. We can, because we know what their epigenomic signatures are as well. Mathematically, from a bioinformatics technical approach, we can parse them out. We can do what is called deconvolution, where we can take out signatures that we know are coming from inflammatory cells, and then look at what is left and see if that is matching up with the tumor.

Dr. Ruff: That is really exciting. In terms of the report that I read, I have seen over the years that the diagnosis term has changed. Early on in training, it was the diffuse midline glioma, H3 K27M mutant grade 4, and now the entire integrated diagnosis appears differently on the reports that I read. Can you speak to that?

Dr. Horbinski: Yes. Well, in the overwhelming majority of cases, it is still that K27M alteration, which, by the way, technically it is K28M. When the original papers came out, they misread the codon, and they thought it was codon 27 when it is actually codon 28.

In some pathology or molecular reports, you will see it correctly reported as K28M. That is the same thing.

There are some other histone alterations that are associated that can, in very rare instances, be associated with this. Really, at this point, the integrated diagnosis, what I typically will do is I will see something under the microscope. "Okay, this looks like a glioma histologically." I might even just release the report on top line as infiltrative glioma, molecular pending. Full integrated report will come out an addendum because that can take two to three weeks sometimes, and so I do not want to hold up the report. At least the physicians and the patients have something to go on for the time being. Then, when the molecular comes back, say, together with the histopathology, the integrated diagnosis is infiltrative glioma, H3 K27-altered, CNS WHO grade 4 according to the new WHO guidelines.

Dr. Ruff: My understanding is that we are also seeing at least the loss of trimethylation and other tumor types, even like meningioma, so it is some common alteration or loss that we find in CNS tumors.

Dr. Horbinski: Yes. There is an immunostain for K27M. In other words, there is an antibody that very specifically binds only to that mutant histone. The one that is known as supposed to be lysine and is mutated to methionine. It is finding that histone. That is a pretty decent, immunostain-specific antibody. It is a way of rapidly screening tissue for the presence of that mutation. If I see that, if I am satisfied it has good nuclear positivity, okay, I can release it as is. I do not typically bother with histone trimethylation immunostain because, as you indicated, that can be seen in a variety of tumors that do not necessarily have, or they are not going to have, these histone alterations. So, the presence of H3 K27 trimethylation loss, and the loss of that particular marker does not equate to this specific midline glioma. You can even see that alteration in certain hematopoietic neoplasms, for example.

Dr. Ruff: That is very interesting. In terms of the patients I see with H3 K27M-altered glioma, generally adult patients do slightly better than pediatric patients. It is a poor prognosis for both, but I counsel them as such. Typically, for children currently, median survival is less than a year. For adults, it is slightly over a year.

I became familiar with dordaviprone several years ago. Probably just after the pandemic. This was primarily in the setting of patients with recurrent midline glioma, H3 K27M altered. There was an expanded access program that we were able to deploy and became familiar with the medication through that; we found it to be remarkably benign, essentially no side effects in my experience. We would always monitor for QTc prolongation with EKGs, but I really never saw it.

More recently, I was the Mayo lead PI for the ACTION study, which is the newly diagnosed setting. Since the accelerated FDA approval, I have been using it in patients with recurrent disease. My experience primarily has been that it is a remarkably benign medication. I feel that it has conferred benefit to my patient population.

One caveat I would say is that probably the patient population that I see benefit more than other populations in that the adult patient population is generally, and Dr. Horbinski, please speak to this, but generally more likely to have the thalamic disease. That seems to be one of the more responsive sites to dordaviprone. These tumors can occur all through the midline, even in the spinal cord. Personally, I would have no hesitancy using any tool I can. That is generally the situation we are in. However, this dordaviprone may be less effective in that population. For instance, the ACTION trial did not include spinal or frank DIPG, so I think probably the patients I see again, adults being more likely to have thalamic supratentorial H3 K27M, they are more likely to respond to the dordaviprone agent.

Dr. Horbinski: One of my suspicions, and this is just a suspicion, is that some of these therapeutics could be effective, but in some sense, if they are too effective, the patient is going to decline and die because if the tumors are - think about this. If the tumors are rising in the pons, that is such a delicate site. If you have a therapeutic that causes massive cell death, or in the case of CAR T cells, cell death with inflammation, well, what is the one thing that an already swollen pons is not going to be able to tolerate? In my opinion, in this particular context, especially, it is not enough to track survival and those sorts of things. We have got to have post-mortem studies integrated into this so that we can get a good, hard look at that tumor tissue to assess the response. For all we know, some of these patients may have responded very well, but still expired early or had a disappointing outcome. Then you look at it under the microscope in the pons; there is hardly any tumor cells left. They are all gone, but the CAR T cells gobble them up, they kill them, but the resultant inflammation was intolerable to the pons, and it caused herniation, it caused dysfunction of the brainstem. If my hypothesis or suspicion is correct, that could explain, to some extent, why the thalamus you see better results there because it is a little more forgiving.

Dr. Ruff: More room. Yes.

Dr. Horbinski: You are right up there next to what is the thalamus butt up against? It butts up against the ventricles.

Like I said, it is just a hypothesis, but I suspect that some of the - because that is such a delicate area, more so than even any other part of the cerebrum when we are talking about glioma therapeutics - even very effective treatments could still result in bad outcomes.

Dr. Ruff: Absolutely. Autonomic function, heart rate, respiration, and so forth live dorsal brainstem, including the pons. Some of my patients will have all sorts of strange autonomic symptoms; orthopnea, kind of an Ondine's curse phenomenon where, with position changes, you stop breathing, and so forth, so all sorts of very delicate circuitry lives there.

Dr. Horbinski, it is a great point. It is rare that I will have patients in my practice undergo autopsies or donate the brain. It comes to mind, actually, because yesterday I had a patient who is doing well saying, "I want to donate my brain to brain tumor research." It is an uncommon thing. Yes, I suppose unless it is built into the trial, you will not get that as much. You will not actually have a final answer. It will always be this assumption that the therapeutic failed, or it did not get where you wanted it, unless you actually prove maybe in life with a window of opportunity, phase zero sort of a thing, or upon death after exposure to the medication. Yes, I totally agree.

Dr. Horbinski: It is understandable in some sense because who wants to talk about that? Who wants to build that into a trial and say, "When this fails, we want to autopsy you"? My experience has been that patients are actually more embracing of this than even the treating physicians. Very often the treating physicians are reluctant to bring it up because it is tantamount to admitting failure. I understand that.

However, the insights that can be generated are invaluable.

Dr. Ruff: I wonder how, because in my world, we are very excited about these new metabolic imaging approaches, so PET imaging for brain tumor. I think that is going to be really interesting to see how frequently we see activity in the brainstem and we do not suspect it, for instance.

Dr. Horbinski: Yes. This is where things got a little tricky because sometimes the patient's last MRI would not happen until like six months prior to death, so the quality of the data was a little shaky. What we were picking up on MRI was a little bit of T2 signal. Not even FLAIR, just a little T2 signal in the brainstem. When we saw it radiologically, that was a pretty solid indicator that, yes, there is real disease down here. One of the things that, like you indicated, would be useful is if we can do more advanced PET and at least try to screen these patients and get a more accurate sense of when the end is likely near. It is not something anybody wants to talk about, but it is useful information for them for end-of-life planning.

Circling back to our original topic here, diffuse midline gliomas, that insight, which in hindsight in some sense feels obvious, but that insight would never have happened had we not been doing these autopsies routinely for other purposes. It was a serendipitous, accidental discovery. I was not expecting it. I had no a priori hypothesis on this. Likewise, with these clinical trials associated with diffuse midline gliomas, in my opinion, most of these patients, unless obviously the family declines, and that is their right, most of these patients should be getting an autopsy to look and exactly see what did that therapeutic do, if anything.

Dr. Ruff: That absolutely makes sense. It reminds me of historically they would say, "There is no need to even biopsy. We know what it is. We know that we can use radiation." Now the move has been absolutely towards biopsy, not only for this specific patient to understand what the diagnosis is and treatment modalities that may best help them, but also to generally help us advance our treatments against this disease. Some of my collaborators here were able to perform a biopsy on a DIPG or H3 K27M-altered tumor and then grow those cells. Generate a meaningful model for high-throughput pharmaceutical testing and so forth. Along the same lines, understanding what happens such that you might be able to intervene upon it, that makes sense. Autopsy makes sense.

Dr. Horbinski: Especially now that we have got so many advanced spatial omics assays that we can put the tissue through, and the fact that these DIPGs, or diffuse midline gliomas, they are not that big. They seem big radiologically. But compared to the rest of the brain, certainly compared to the rest of the body, they are not that large. What that means is I can get you very, very, very good representation of that entire tumor on one slide, which is not necessarily the case for a lot of these other tumors that are crossing the midline and got the butterfly lesions and all that. You can have all kinds of different subclones of glioblastoma, for example, spreading throughout the brain. DIPGs, DMGs are not like that.

Just one really good tissue section, you can do multiplexing to assess very rigorously all the different immune cell populations there. There is advanced multiplex metabolomics that can be done now, transcriptomics on fresh tissue. There is a lot of really, really advanced things that we can do to squeeze even more information out of this tissue than we ever did before.

Dr. Ruff: That is very interesting. I had not thought about that with glioblastoma. It is such a heterogeneous tumor. There is this headline that you can have tumor cells from different areas of the tumor that have less in common with one another than maybe a tumor in someone else's brain. Just that level of variation in gene expression. That is not the case with these. They are fairly uniform throughout the tumor as we understand it.

Dr. Horbinski: It is such a small space. The tumor does not have as much time, or does not need as much time to grow and evolve before it is going to cause some symptoms Some of these other tumors, depending on where they are at in the cerebrum, can be there for quite a long time. Not just gliomas, but meningiomas and these other tumors can be there for a very, very long time. Then somebody's got some symptoms. They got headaches. You scan them, you are like, “How did this tumor not manifest clinically sooner?" That is not the case with these midline gliomas.

Dr. Ruff: Absolutely. Yes, it can be a very rapid clinical course. For instance, often the history that I will hear is some bulbar symptom and double vision, maybe dysphagia, slurred speech, something like that, or maybe facial numbness, for instance, trigeminal involvement, which will eventually lead to an MRI scan. Typically, here neurologists will see the patient first, then followed by a neurosurgeon. Of course, we are relying upon radiographic imaging to help give us a sense of what this may be, which is imperfect as we mentioned in prior cases, and often there is multidisciplinary collaboration in terms of I will speak with the neuroradiologist, say, "Give me your best guess of what this represents." Then, very likely, I may pick a specific surgical colleague and ask them, "How comfortable are you with the biopsy?" Then we proceed with tissue diagnosis.

Typically, for H3 K27M-altered standard of care at our institution would be that of upfront radiation. Typically, we do not use alkylating chemotherapy. We do not feel there is tremendous benefit to that. Then we quickly look for any evidence of recurrence. If there is any sense that maybe recurrent disease after radiation will initiate a dordaviprone, fairly low threshold to do so, or alternatively, look for trial options for these patients. That is my typical workflow for these folks.

Dr. Horbinski: Yes. I think that is entirely appropriate, especially trying to get people on as many trials as possible. That is the first thing because even the new therapeutic approach, dordaviprone, in particular, is really interesting. It is exciting. It is not curative, so there is still room for improvement.

Dr. Ruff: Yes.

Dr. Horbinski: The alkylating chemotherapeutics, in particular, in my opinion, nobody who has a DMG with K27M should be on temozolomide or any other alkylating chemotherapeutic because of two reasons. First, as you indicated, the trials have been done ad nauseam, and they just have shown no benefit. In part, it is because these tumors in my experience never have MGMT promoter methylation. To really see a meaningful benefit from temozolomide, you want to see that methylation because that turns the MGMT repair gene off, and so the tumor cells have a harder time repairing the damage done by temozolomide. Yes, the fact that you never see that methylation, that means that gene is always turned on in these DMGs. It is just not useful. It is not a useful therapeutic strategy.

Dr. Ruff: Yes. There is tremendous therapeutic desperation in these cases, in my experience, and so I certainly see it used temozolomide, alkylators, and so forth, but yes, I totally agree. I do not advocate for it in our folks. With dordaviprone, I think the mechanism of action is hard for me to explain to people. It does not directly target the alteration; I generally think of it as more of a mitochondrial poison for tumor cells, preferably, which then leads to cell death pathway. However, the actual mechanism of action may be more pleiotropic or maybe not entirely understood. It is a remarkably benign medicine in my experience. Probably the most benign medicine I use routinely, more so even than antiepileptics. A lot of antiepileptics I use, folks have all kinds of side effects, whereas this, folks generally do not.

Dr. Horbinski: Yes, it is an interesting therapeutic. It has multiple effects within the tumor cells, which may actually help explain why you see at least some therapeutic benefit the general rule of thumb now is in cancer, if you have a drug that blocks one pathway and one pathway only, you may or may not see good response because there is just so many other ways to get around it. Now, there are notable exceptions. Like we have the new KRAS inhibitor in pancreatic cancer that is significantly extending survival, which is magnificent, but even then, that is not curative. If we are really going to get cures, it is going to have to be multi-armed approaches or drugs that have more than one mechanism of activity.

Dr. Ruff: Absolutely. Yes, I totally agree. Well, Dr. Horbinski, this is great. It has been a fantastic discussion. I do not know if there are other elements that you would like to hit on.

Dr. Horbinski: I think, at least from a pathology standpoint, we are all the time now working on really advanced next-generation molecular assays to even further provide even more information for the patient, for the treating physicians using even less material than ever. The things that we are now able to do, even my laboratory and on a collaborative basis, the things that we are advancing into clinical use are, I hope, going to be revolutionary. They are going to generate a lot more information and much more cost-effective as well. As is the case in a lot of fields, the technology drives the advancements. In this case, the advancements that we are seeing in molecular biology and molecular testing and assays are really, really helping us.

Dr. Ruff: Absolutely. It has been very exciting over the last eight years to see all sorts of different targeted agents come about.

Dr. Horbinski: Yes. More reliably and more rapidly classify these diseases using less tissue or in some cases, liquid biopsies. Yes, there is a lot of things that I think are going to be coming down the pipe, especially once we start getting artificial intelligence involved in really doing a deep dive into the analysis of this multi-dimensional data that we are generating from these assays coming to coming up with, for example, integrating in a more rigorous fashion, molecular results with the histopathology and the radiology and the overall patient's profile to generate probability-weighted differential diagnoses for a lot of these tumors. These are the things that are coming along.

Dr. Ruff: Absolutely. Yes. I am hopeful for disease monitoring as well. Not just the initial diagnosis, but like ongoing response assessments and so forth, not just relying upon neuroimaging, but having CSF or whatever it may be, cell-free DNA in the blood, whatever it may be. That would be amazing. We would love that.

Dr. Horbinski: Exactly.

Dr. Ruff: Yes. That is fantastic.

Kendall Schick: Thank you so much, Drs. Horbinski and Ruff. One of the things that you both touched on was how to bridge that communication gap with not only patients, but also practitioners. For this program, we are also partnering with the NBTS, and they have the MyTumorID, which allows for patients and providers to access practical evidence-based resources, including materials for clinicians that they can share with patients to support these informed conversations about biomarker testing and personalized care. It is just a very exciting space, very exciting time.

Dr. Horbinski: Yes, I think pursuing just as a follow-up to that - there could be a whole other conversation - one of the limiting factors is not even so much the technology anymore; it is the adoption of the technology on a routine basis. A lot of our patients are still being diagnosed with primary diagnosis happening in the community setting, which is fine, but what I see all the time is then when the patient goes to Mayo Clinic or previously Northwestern or whatever, their tissue comes with them, and I will look at this, and it will be like, "We have got a diagnosis of glioma here. We need to subtype it by molecular diagnostics." There is no tissue left because the small biopsy that we had in the first place was wasted with a lot of immunostains that were unhelpful that should never have been done. Not only is it a waste of money, but more importantly, it is a waste of tissue.

What I would really like is for any even suspected glioma, just to reflex, automatically send it to a quaternary care facility, where you have got people who do nothing but that for a living, and can look at it very carefully and say, "All right, here is how we are going to approach this case," without wasting any tissue at all, because otherwise, the patient's got to go in for another surgery. That is a failure. It is a failure of the system.

Dr. Ruff: Absolutely.