Good afternoon, and thanks for braving the weather and the attrition as the day has gone on from people leaving the meeting, so I hope everyone's had a great meeting, and thank you so much for coming to our symposium. We're looking at molecular profiling and gynecologic cancers, now and into the future. I'm Dr. Tom Herzog, and I'm really honored to be joined by such a great panel this afternoon, and on the far side of the stage there for me is Dr. Jubilee Brown, in the middle, Dr. Shannon Weston, and Dr. Matt Powell, so I'm really excited to get this going. These are my disclosures. So this whole idea of precision medicine has been one that's been around for quite a while, and it really has taken a little bit longer in GYN cancers than it has in some other cancers to really make a difference in terms of treating patients differently. So we think of where we were and maybe where we are in some areas and some subsets of our diseases, we still treat people pretty much the same, irrespective of their molecular makeup, and we've certainly had a great amount of discovery in terms of prognostic markers, but we're really moving into the year where we have great examples of predictive markers in our therapies now, and so the push, if you will, to do molecular profiling and better understand our tumors is greater than it ever has been. Precision medicine, everyone's seen that, but it's really about putting the patient in the middle and working these other elements and platforms around them so they get the right medicine for the right tumor. Many things have happened that have allowed this to occur in terms of things we may not even think about in terms of things like data storage, for example, and cost of sequencing, some other things I'll address real briefly, but there's been a lot of breakthroughs in a number of these areas beyond just molecular genetics that have allowed this to become a real revolution in terms of influencing our therapies. This is a timeline of some of the major achievements in sequencing technologies, went all the way back to the late 70s with the publication of the Sanger sequencing, and then looking at PCR that came out in the early 1980s, and then the commission of the Human Genome Project that took place, took 10 years, $3 billion to do. So a lot of money spent on that, and at the very end I'll give you an idea of how technology continues to improve in bringing down the cost of sequencing. The TCGA pilot project was launched in 2006, and then we had the publication that came out in 2013, and you can see we continue to move on. I've got some other timelines here, and you can just see the interest of this whole concept of mapping our biologic information to develop therapeutics, and I think that it really is an exciting time to be in medicine, and we have opportunities ahead, and I think for people that are in med school now and moving forward, it's going to be a completely different world than what we faced, which is a great thing. I think you're going to have so many more opportunities for different types of treatments that will be much more specific, hopefully less toxic and more effective. So another look at this, sort of looking at some of the technology things there on the bottom and some of the actual things that came out that were introduced on the top, but again sort of similar to what I just showed you. The same type of thing here, looking at the different mutations that were interrogated on the top, and the different approvals that occurred on the bottom. So again, lots of activity in this area in really the last decade, it's just been transformative. So I think in terms of really taking care of our gynecologic cancer patients, what tools do you need to best be able to take care of them? And I would argue next-gen sequencing is probably requisite in most cases, but we're now finding you also need to understand what's going on at the RNA level, you need to understand what's going on at the protein level, so IHC comes into vogue as well, and you can see all the different drugs that fit into the choices there based on the platform that you're interrogating, whether that be IHC or RNA or DNA sequencing. So we're going to be talking a little bit about DNA and whole exome sequencing, RNA sequencing and protein sequencing. This particular symposium is sponsored by Keras, and so we'll be talking about some of the things that they bring to the market, which I think there are some differentiators from the competitive landscape in terms of depth of coverage, micro-dissection, and some of the other things that they do so well. What about ovarian cancer? And I don't want to get into anyone's talk here, but I think the first foray into precision medicine is really recognizing that different histologic subtypes really were associated with different genomic alterations, and that's what we see here. And now we lead to, with the dawning of the PARP inhibitor era, an understanding of the importance of loss of heterozygosity testing and determining homologous recombination deficiency so that we can best choose who goes on to one of these types of therapeutics. This is Keras' genomic LOH that's been out now for about a year, a little less. And I think that, you know, from the standpoint of the technology, it's industry standard in the sense of 250,000 SNPs with one megabase resolution. The question is, you know, how does this compare to FDA-approved Myriad MyChoice and Foundations companion diagnostic? And there's efforts to validate this, and we'll be seeing more data from this as we move forward. So it really is exciting to see this type of thing put into our report as well. This is just looking at it again, and sort of gives you a little more background on it. They use a cut score for LOH of greater than or equal to 16 percent, much like Foundation 1. So I want to introduce our first speaker. I believe, Matt, you're up next, right? So Dr. Powell is the division director and professor at the Washington University in St. Louis. Matt? Thank you, Tom. And they're clapping for you. They're clapping for me. Thank you, Tom. And I know it's been an exciting meeting when it comes to molecular profiling for endometrial cancer. And two days ago, we had good controversy and a nice master's session in endometrial cancer. And I think if you haven't thought about molecular testing in endometrial cancer, maybe you weren't at that session. But we'll try to go through this and try to highlight some of the practical aspects of testing and why we should be considering this. Here are my disclosures. Practical testing, what is needed? What are our guidelines telling us? You've seen NCCN, and we'll be reviewing a little bit about the WHO and ESGO guidelines. And how do we separate out prognostic markers from truly predictive markers that should influence the way we practice? And how do we get timely results, make this cost efficient, easy to use reports? We've known about mismatch repair IHC testing for about 20 years, and I would say it took a good 10 years to get people to routinely implement this in their practice. We want to minimize the hassle factor. We want to minimize those phone calls for those peer-to-peer calls if you have a denial, and really work with our pathologists to have their workload manageable. I'm not sure how many of you have that, gee, I really wanted HER2 testing, I have to send an email, it has to be HIPAA-compliant email, I have to figure out how to get that to work within my health system, and it still doesn't happen, so you don't have a good triage mechanism to how to get that back to yourself. So dealing with your pathologists, I know we all get a favorite pathologist at our institution, is one of our big challenges, because for whatever we do, it requires that partnership with our pathologists. When to perform the test, and we'll talk a little bit about this, can we do this on our preoperative or pre-hysterectomy biopsy, or should we wait for the final specimen? What is that best specimen, and how do we assure quality? This is a timeline of the last century plus in the management of endometrial cancer, and I think we spent most of the 20th century doing hysterectomy and adding radiation to most every patient, and for centers that had radium, that was the big thing going on in the late 1910s, everybody got radiation, we didn't know who needed it, who didn't. Endometrial cancer was a relatively rare cancer for most of the century, and it's really come on now in the 2000s. In the late 80s, we learned about lymph node involvement of endometrial cancer and found that to be prognostic, and so rather than studying that in a rigorous fashion, we jumped on that bandwagon saying, well, if it works in cervical cancer, helps us predict who needs therapy, maybe it does in endometrial cancer, which led to finally some randomized trials with Aztec and the Italian trial, which really questioned the true value of knowing the nodal status, sentinel lymph nodes became our surrogate. Along this time, we started to deal with fertility sparing procedures, the genetics of endometrial cancer, and a lot of work, and probably over 2,500 publications talked to us about prognostic markers in endometrial cancer. And for those of you that are involved in fellow education and fellow boards list presentations, I would say a good portion of what we saw come to their ABOG exam was a prognostic marker in one of our cancers that we'll be talking about today. But how many have actually made it to be a predictive marker? And we'll talk a little bit about that. So the TCGA now, I think we'll show those slides, really called out that endometrial cancer's not one cancer, four main groups, and we'll talk about those groups, and we'll probably be refining that into eight or nine groups here soon. The amount of lymph vascular space invasion, LVSI, I think Portec's done a good job in compelling us to think about this as a three-tiered system. I'm not going to show the data there, but think about extensive LVSI being what matters, limited, probably not so important, but we should be reporting that as such. And hopefully we'll be seeing that through our work through AJCC and CAP, whoops, sorry, that we'll actually be doing this breakdown in lymph vascular space invasion. Now how do you prove that that's a helpful intervention? Well, it takes time, and we'll probably have to collect that for a decade before we know and can see that from databases. So I know we all want the answers to many of these questions now, but it does take time. So P53 mutation, as we mentioned, that's part of identifying one of the TCGA subgroups, but also may be helpful in predicting who benefits from bevacizumab from our GOG86P trial, and I'll show you a little bit about that. Portec-3, making sense of their molecular analyses. We still don't have our molecular data for 258, but when there wasn't an advantage of radiation therapy on that trial, it'll be hard to show if there are maybe some subgroups that potentially would benefit from the role of radiation. We'll be talking a lot about poly-de-escalation today, and I think that's one of the big factors that keeps many of us from testing for full molecular testing in endometrial cancer is how do I do poly-testing in my practice, and we'll be touching on that. So predictive biomarkers, Portec-4a was envisioned now over a decade ago. This trial has accrued and really looked at managing early-stage endometrial cancer patients, placing them into a favorable population, intermediate population, and unfavorable. This was launched really just as TCGA was coming out, recognizing that poly was an important favorable prognostic marker and likely predictive marker of somebody that didn't need therapy, and then also the MMR intact CTNNB1, or continued non-mutant patient without lymph vascular space invasion. You could really figure out that's a population of patients that didn't need treatment. Across the top here is the markers that we're looking at, and those include P53, our mismatch repair genes, something called L1-CAM, which is kind of a finicky little marker that usually P53 and L1-CAM are fairly collinear. ER is also something that many of us are advocating testing because we can further subdivide our risk groups by looking at estrogen receptor. There's a nice poster that was up yesterday looking at that from the Canadian group, showing how ER may further refine some of our prediction, and we talked about LVSI quantification. So if the patients were found to be favorable, they could be observed. If they were intermediate risk, as in they have a CTNN mutation or mismatch repair deficiency, those patients were targeted for vaginal BRACI therapy, and if you were unfavorable, as in a P53 mutant, which was MMR intact, or you had substantial LVI or this elevated L1-CAM, those patients were targeted for pelvic radiation. It was estimated you could dramatically decrease the amount of pelvic radiation by using this trial strategy, and again, we don't have the results of this yet, but they've fully accrued the patients, and hopefully we'll have some prospective data in the next year or so, or more, helping us further de-escalate strategies. As I mentioned, the TCH subclassifications, I think this slide's probably been shown about 80 times in this conference, and really breaking these out in groups, and it's important to understand the stepwise process of how we look at this. So POLI happens first, because POLI causes such a mutational phenotype, you start to see MSI happening, you start to see other things, as in P53 mutations happening, as a bystander effect rather than the driver. So that's why we look at P53 first, we do MSI next, and then moving on, and you've heard some controversy about these double classifiers. What does that mean? I think Dr. McAlpine and the Canadian group have done a good job looking at a fair number of these patients that have double classification, and they're able to say, if it's POLI-driven, they're probably going to behave like a POLI tumor, and those double classifiers really fit this logical stepwise mechanism of calling it a POLI tumor first if it has a POLI mutation. Now the other controversy about POLI is you need to make sure it's actually truly a true mutation, and I'll get into that here in a bit as we talk about some of the data out there from POLI, because there is some controversy there. When we look at, really, the bad actors, are the serous-like, or P53 mutated, or copy number high subgroup, and that's the group that really causes the most havoc. I mentioned a little bit about the guidelines, and I think probably the ESGO, ESTRO, ESP, or Spanish Guideline group are the first big step, and you're going to see this happen with kind of all of our staging. You'll see within AJCC and PHEGO, dramatic changes are coming. I can't share those yet, but in short order, you'll see something that looks fairly similar to this for our other staging subgroups, and the one looking at low-risk groups, obviously if you have molecular classification known versus unknown, so the unknown is shown on the middle column there, and then if you do know the molecular stratification, it's falling on the far right column, and you see even stage two POLI mutated tumors with no residual disease felt to be a low-risk situation where that patient could be observed, and I know that's a big step for many of us, and that's one of the main reasons driving molecular testing, and we might ask, do we have enough information to de-escalate to that amount for a stage two tumor to not receive additional therapy because it's considered low-risk, and we'll talk a little bit more about that. The cost here, obviously P53 and other factors can drive our patient into a high-intermediate and a high-risk population. Still exactly how to triage these patients and really what is their best adjuvant therapy still remains unknown and is being discussed. Testing, WHO now does recommend integration of microscopic features with molecular characteristics as the best approach to stratify patients to predict prognosis, and to do this practically, so we have the World Health Organization saying we should be doing this. So POLI and mismatch repair deficient tumors really can occur in almost any type of endometrial cancer, usually not in our serous type cancers, but we'll talk a little bit about that as well. Tumor disproportionately tend to look higher grade, and they tend to have a lot of heterogeneity within the tumor, so you'll see different parts of the tumor expressing different features, and you can imagine a tumor that has a high mutational phenotype obviously looking more heterogeneity, and often can have some of this histologic ambiguity, so the tumor looks a little funny. Sometimes our pathologists are able to determine these even before doing molecular testing, but still certainly molecular testing is required. So which cases do not require POLI testing? People have looked at this in a couple different ways, but I would say if you have a grade one or two endometrioid carcinoma that's stage 1A, there's no or focal LVI, maybe one focus, it's ER positive, P53 wild type, and MMR proficient, probably don't need to be spending the time doing POLI. It's a low risk patient population, it's very unlikely to be POLI, and even if it was POLI it's not going to change your management because you're going to observe this patient very likely given these criteria. So as we come to triaging, we can use hysterectomy information as a stage 1A low grade patient and not perform molecular testing in that population, and you'll see many guidelines that try to accomplish that. As I talked about POLI variants, and the POLI, the exonuclease domain, we've identified what we truly think are the important key mutations, and I think this gene is subject to mutation in cancers, and a lot of these don't seem to be causing the effect that we see with this proofreading problem where we get this massive amount of neoantigens and mutations happening that we do when we see these mutations within exons 9 through 14. And there is a specific germline POLI mutation that the C1270C to G that should hopefully be tagged. If you see this one, you may want to think about referral to clinical genetics because this is one that can be a germline problem. And this I think was a bit controversial within the master session, and this was a bit of a debate between Dr. McAlpine and Dr. Latteo, Memorial has a single institution publication about their POLIs where it was quoting a 17% recurrence risk. Within that publication, they did their in-house testing, and for patients that had recurrent disease, they added those into their population. So I don't think you can truly call it a 17% recurrence risk because it was artificially inflated because when patients had recurrence, they were referred back. Long story to that. But there was also some nonpathogenic POLI mutations within that. So what this meta-analysis had done is carefully look through the literature and make sure it truly was a known mutation, an inactivating mutation, and looking at the therapy received, and what can we say about these patients that have POLI mutations. I think they really concluded patients did not appear to benefit from adjuvant therapy, and they routinely do very, very well. So is this enough evidence to say, POLI, leave that patient alone, it's going to do well? And most of us would say no, but I'd liken this a little bit to the management of low-grade serous ovarian carcinoma. If you imagine back in the 2000s, Dr. Herzog and I did a debate, it was about stage 3 low-grade serous borderline tumor of the ovary, chemo or not. Every one of us decided now that that patient doesn't need chemotherapy, but back then it was still controversial a little bit. But major thought leaders in the field said, you know, these patients do routinely well, chemotherapy doesn't seem to benefit them, and I think we should observe them. That was a big change, but it did take data to show these patients did well, but now we all routinely accept that we don't treat stage three borderline tumor of the ovary with chemotherapy, and that was probably based on less data than we have now. Because these patients do not appear to benefit from adjuvant therapy, but we want to prove that, and we'll get to that in here in a second. So when you look at your TCGA subgroups across different populations, what do you expect to see? And this is a very nice review that was in Nature in 2019, and I want to call out a few things. In the patients kind of with the best prognosis are shown on the left with poly mutations, and then we have our copy number low group. Now, not all these are exactly the same across here, but when you look at a younger patient population, you're gonna see a little more poly, and I think that's important as we think about fertility sparing, that sometimes these patients will have poly mutations, and you're gonna see less P53, and I think we've all known that from our practice. One of the reasons GOG99 had age as a risk criteria was probably this factor right here. We see less mismatch repair deficiency. Again, there's less promoter methylation of MLH1 and less P53 the younger the patient is. So that surrogate of age is really just a surrogate of P53 and mismatch repair deficiency. I probably won't go into this much more detail, but it's a nice look at how different populations can show different risk patterns of their TCGA subgroups. Collecting a specimen, we know we can do molecular testing on preoperative biopsies, PIPEL, endometrial biopsy, DNC. Sometimes these are better preserved specimens than the final hysterectomy specimen, so that there is good evidence that molecular testing can be done on these. And when I looked, and I mentioned before about testing, I said which groups do not require polytesting, I showed here, and then which do. Kind of basically anything if you're not in that top subgroup. The British Association of Gynecologic Pathologists have their recommendations out at this point and basically summarizes this in a little pragmatic way here, just showing a stepwise process when you look at testing. And I think that if many of you are moving towards testing on your preoperative biopsy, you won't have much of the information, but moving ahead with testing early is a good way to do that and getting that information. It's surprising in the U.S., I think we're quite behind many of our European colleagues in the rest of the world and certainly behind Canada, where we see a lot of this already happening. So efficient testing, I'm gonna skip through this slide, but selective molecular testing has been reported by this Italian group where they did MMR analysis on all their patients, much like we would do, and if it's early stage, you'd move ahead with P53 analysis, if it's advanced stage, they said, well, you're gonna treat that patient anyway, so we don't need to move ahead with that. And then if it had any of the further features, they would move ahead to polytesting. And by utilizing this strategy, which again, takes quite a bit of involvement from your pathologist, you would have a novel number gram that polytesting was decreased by 67% of groups and P53 analysis by 27%. So using a step-wide process means you wouldn't have to do everything. But as I mentioned, many of us don't like the hassle factor of having to reorder tests and track down things. So really, the need for an assay that can handle all that for us, and there are several commercial assays, I'm not listing all of them, but many of our pathologists have preferred assays or in-house labs that try to allow us to get polytesting. We've gone to using Keras mostly because we like the ability to have both immunohistochemistry and NGS and a workable report in a one-stop shop. And I think that's been our biggest issue, is trying to get all the information we need in one spot at one time and with good turnaround time. So many of you that have used the assay know a quick two to, usually a two-week turnaround time. Also, the hassle factor seems to be very manageable. As we make big changes in our practice, I know our payers aren't used to looking at early-stage endometrial cancer, but we're quite happy that we're allowed this to be covered and minimize those peer-to-peer encounters. So what do these cancers look like? And I think this is a nice, from the IMPACT assay at Memorial Sloan Kettering. This is the TCJ subgroups of 1,900 of their patients. I just wanted to focus a little bit on poly, what do those patients look like? The majority of them are endometrioid, but they're of equal grades, grade one, two, and three. And so you can't really, that microscopic view doesn't really help you. But even in their population, obviously the poly patients do really well, and it's that copy number high group that we need to focus our energy on. And we've done this through the NRG. So our MSI high-positive, high-intermediate-risk patients, they were enrolled on a study called GY020. This is plus-minus checkpoint inhibition. We've accrued all those patients, and hopefully, in short order, we'll have results of this study for you to let us know that we can use a year of checkpoint inhibitor for our high-intermediate-risk patient population to hopefully help improve their lives. For our HER2 population, we know that Dr. Madden-Nichols-Fader's trastuzumab study did show, in a randomized phase two fashion, improvement in progression-free and overall survival for the patients. And now we have implemented GY026, which is pertuzumab plus trastuzumab. And then I'll be talking to you about what's called the TAPR trial. So we're hoping to open this trial soon. It's a Canadian trial through the CCTG, which really is to prove that we can de-escalate care for our patients with POLE mutation and the P53 wild-type NSMP group. And again, 120 patients with POLE will be followed, and 180 with the P53 wild-type NSMP group. There will also be a higher-risk exploratory cohort of 25 patients with stage three, or higher-risk POLE mutations, and these patients would be followed carefully. Is it ethical to enroll a stage three patient on an observation trial with our current level of data? And I would argue it is. It is up to the patient in a discussion. But part of that conversation has to be, we think we have a back-pocket solution for that patient. Immunotherapy seems to work quite well for these very high tumor mutational burden patients. So even if those patients do come back, I think we have a salvage mechanism that should be quite high. So we will learn something from this trial, but it is, we'll take some forward thinking to proceed. So when we look across molecular groups, we think the MMRD population probably does not benefit from chemotherapy. The P53 abnormal group probably does benefit from chemotherapy. That magnitude of benefit, though, is a bit unknown, and we'll hopefully learn from some of our upcoming trials about that. NSMP group, probably the ER positive group does well. We need to develop strategies for our ER negative group, and then we've talked enough about POLI. Just a quick point about rebiopsy. I do think we should be rebiopsying our patients, as we sometimes will see mismatch repair deficiency come back in the recurrent tumor versus the primary. And a lot of this just has to do with some of the tumor heterogeneity. PD-L1's probably not needed in endometrial cancer. It does not seem to be predictive. And we have some data that depending on their type of mismatch repair deficiency and response to checkpoint inhibition, it looks like those that were promoter methylation maybe didn't do as well across two different studies. But as you saw from the DistartelMAP data, maybe that's not the case in a prospective manner. Just looking at our upcoming GY-026 trial that's now open. I encourage everybody to open this at their institution. Looking at further intensification of her two targeted strategies in endometrial cancer. I did mention this briefly on the overview about the role of mutant P53 in endometrial cancer in targeting those patients with bevacizumab. Again, this is the publication. The initial publication was Dr. Engajanian and then the molecular, Kim Leslie, showing this mutant P53 seeming to do better with addition of bevacizumab. So we have some new markers. I didn't talk about selonexor, but maybe there's something special for our P53 wild-type population. But molecular testing in endometrial cancer, I think, is here. I think now it's just time for all of us to operationalize this into our practice. So thank you. Thank you. Thanks, Matt. We have a couple minutes left in this part of the session. We'd ask anyone from the audience to please go up to the mics or just scream out your question if you have any. Just raise your hand or stand up so we know you have one. While we're waiting to see if you have any questions, let me start with this, Matt. What did you think of the presentation and the first plenary? It was Amy Jamieson, I think, that presented it, that looked at taking the PROMIS approach stepwise and doing it all at one time. Yeah, so the PROMIS 2 classifier, basically a one-stop shop. And I agree, I think the amount of headaches we already have in medicine are enough. And if we can have a single, stand-alone assay that allows us to do that, certainly makes our life better. And should it just be limited to what we normally collect or just the four PROMIS classifiers? I think that's one of the things I like about a commercial assay like KRAS. I can get as much information, start to see other pathways. That may not help us right now, but as we continue to do drug discovery, drug development, hopefully in the future we'll have some ideas to how best to help our patients. Great. Any other? Oh, there's a great question here. Hey, there, Dr. Chase. Yeah, what do you do then? Right, so yeah, good question. So I know that sparked a little bit of angst in the master session a little bit. It's probably a little early to modify our surgical approach yet. I don't know the answer. It's kind of, do you abandon doing lymph nodes for serous borderline tumors? And many people have, for that reason they don't seem to matter anymore, right? So we may see that, but it's gonna take a lot more time. Right now I would still say, especially with the low morbidity of sentinel lymph nodes, I would agree still collecting that would make sense to do. Would you ever use COLE at all? I mean, I think you very well could. I mean, I'm advocating that, and again, WHO is advocating it. You're gonna see that from many different groups that we should be collecting this information and using that probably to the best of our ability at this point. I mean, will it be definitive? Probably not for a number of years yet. Dr. Nauman. Why do you think that there's a difference maybe? I would think it's tumor heterogeneity. If you didn't hear the question about MLH1 promoter methylation patients versus somatic or germline population, you know, we got really excited about the distarlomab rectal cancer data where they had nearly 100% response in rectal cancer. And those patients are primarily all germline somatic population. You don't tend to see a lot of promoter methylation in rectal cancer. Now, interestingly, with that update of the distarlomab trial in endometrial cancer, I know it was only 9% of the population that were probably not promoter methylation or somatic germline. They didn't seem to do markedly better, but it was a small group of patients. The emerging data. So either there's more tumor heterogeneity, as you point out, and they're progressing on those patients that are somehow maybe don't have the higher new antigen load. I don't know yet, but it's an interesting surrogate. Well, and I think that's one of the tricky parts. When those patients progress, would Lempembe be the appropriate next regimen? But unfortunately, right now, it's not approved. But they can go on GYO25, which is ipinivo versus nivolumab. So think about that in the future if you have somebody that recurs and you're interested. You know, maybe single agent checkpoint's not enough for some of these patients. Matt, can I ask, what do you think about sending biopsy specimens before hysterectomy? Right. Yeah, I mean, I think that's part of the idea of the PROMIS-II is actually sending that preoperative biopsy. That's where you get all your information you need in one stop. And I think that makes a lot of sense. You may be ordering a few more than you need, especially, you know, that 1A, grade one, low-risk population. But from a looking to the future, I think that's probably the way we'll be going. But I think that's kind of getting at what Dr. Chase was asking. Because if you don't know the status of lymph nodes, let's say, so say you do de-escalate, then is your poly patient a stage one or two? Or is it a stage three or four? So I think, you know, for those of us that do sentinel mapping, and I found it very interesting hearing from our Canadian colleagues that the molecular testing was cheaper for them than sentinel mapping, which is exactly the opposite of what, at least in our institution. So, you know, to us, we do sentinel mapping on every endometrial patient, you know, so we're gonna be able to get that assessment fairly quickly and then utilize that. So it's hard to see right now the utility of using it to direct surgery, but certainly for adjuvant, it's really nice to have it already done. Well, and I think in Canada, where you have a lot of general GYN doing the hysterectomy and in vitro cancer, I think there's probably even a greater importance, not so much from the poly side, but the P53 side to get those patients really triaged to getting their surgery done somewhere where there is more of an assessment of metastatic disease. Yeah, I really liked the Italian study that you highlighted, because I like the idea of de-escalating using poly, you know, not having to necessarily do next-gen sequencing on every single patient if you have to use separate platforms. Okay, great discussion, Matt. You covered a lot of material and testing, and I want to thank you for doing that. Our next speaker is Dr. Shannon Weston. She's professor at MD Anderson and is director of early drug development in phase one. And as I understand, when the clock flipped, you're now the new medical director of GYN oncology. So congratulations, I think. It's probably a lot of work. That's exactly what all my colleagues said. Congratulations, question mark? So yes, thank you so much for the kind introduction and thank you all for being here. I'm gonna spend the next few minutes taking a little bit of a step back and looking at how we've assessed molecular profiling as a predictive marker for clinical trials. So, you know, first we'll talk a little bit about the PROMIS, although Dr. Herzog did a great job of reviewing that. I'm gonna talk to you about some of the early precision medicine trials and what we learned from the way those were done and how they were maybe less successful than they could have been, and then our ongoing successes in the future. And you saw this, this is a slide that we all share, but the goal of precision medicine is moving things forward, right? And so how can we truly personalize cancer therapy for patients? And getting beyond this idea of prognostic markers, I think Annie Ellis today said in this first session, like, we don't need to know cancer patients, patients with cancer do bad. Stop telling us this. You know, let's use markers to actually tell a patient that they're gonna do better because we're gonna treat them appropriately. And I think that's really important. And of course, we don't really talk about this that much, and I'm not gonna talk about it too much, but predicting not only sensitivity and resistance, but also maybe predicting adverse events, too, understanding who might be more sensitive to a given therapy. But all this comes down to is biomarkers, and this is really where our development needs to be. And, you know, I think we are doing a better job. We're moving away from lumping. We're transitioning into splitting based on the aberrations and the testing that we see. And I think, unfortunately, we're doing good at splitting, but we're not necessarily changing therapy, right? So we're still tending to treat everyone the same, although I think we're making progress. So, you know, really what we need to do is determine which of these aberrations in these different tumors are important, right, which are actionable, meaning can we actually target them, and then how is the best way to target them? And so I think that really, as we've seen, more and more tests, as well as more and more therapies get approved, you can see us refining our techniques, refining the way we're doing these trials, so that we truly, at the end of the trial, have an answer that we can now put into practice with our patients. And, you know, Tom touched on this, and I think that it's informative to look at this and see what are we actually getting done? And I think my colleagues have already said, it's really nice to be able to just send one thing and get all of the things you need, because these approvals and these different therapies are really dependent on different tests, including, you know, DNA sequencing, RNA sequencing, protein, immunohistochemistry, and so really understanding which of these tests are important and being able to put them into practice. So having a test that really allows you to assess all those things is important. Okay, so before I get into the details, I want to just make sure to level set and make sure we're all on the same page when we throw around some of the kind of novel trial designs that have been utilized in kind of molecularly targeted therapy. So first are umbrella studies. So umbrella studies are all the same cancer. You get different tests on the same cancer, so lung, these are little, are they livers? They're lungs, lungs hanging there. And then you give a different drug based on the final test result. So for example, the iSpy neoadjuvant breast cancer study, really nice study, they utilized the on bio, or a pre-therapy biopsy to determine EGFR status or HER2 status or what have you, and then bend the patient into a given arm based on those results. And this is a pretty cool platform study where they actually have the ability to rotate arms in and out, and that's why you just see Investigational Agent A, B, C, because they made it a really adaptive design so that as new drugs come to market or come in to be able to be assessed, they can adjust and add another arm. This is a little different from basket studies. So basket studies are all the same drug, but being utilized across a number of different tumor types based on the result of the test, okay? And for example, this was a phase one study of an AKT inhibitor, AZD5363, where they had some suppositions about what might lead to responses to this drug, including the presence of PI3 kinase mutation or the presence of an AKT mutation. And you can see they set up arms based on, one, the presence of that mutation, and then also teased it out by cancer type. So it was breast cancer as well as gynecologic cancers. Early tumor agnostic basket trials were really all over the place. And that's because, again, it wasn't being, it was being driven by what was available, right? So what tests could we do, and what drugs were available? And so we saw some mixed results early on. You know, there was successful testing. If you look at this table, the second line in, you know, 80, 90, 100% could have testing done. So that's good. About 40% or so had something actionable, something targetable. That expanded if they included transcriptomics. So looking at RNA, that expanded up to much higher, like 60 to 70%. That next column is kind of the saddest column because it says who actually got a matched therapy. You can see it was about maybe a quarter of the patients. So they had testing, they had something identified, and yet they still weren't actually treated with a targeted therapy directed towards their results. But when they were, they did see response, okay? So some of the response rates are low, again, depending on the drug that was being assessed, but some were upwards of 25%. In a heavily pretreated patient population, I think we all agree that's a win, although of course we always want to do better. So I'm gonna show you two of these trials that were listed in that table, and just the contradictory results that we saw that made things quite confusing, but also again, just to inform kind of how we should design these trials. So this is the SHEVA trial. It was over a couple of years. They had 741 patient solid tumors and 10 therapies. But I want you to notice that most of them are single agent, and a lot of them have not moved forward for development because of lack of activity. But this is what they had available at the time. 40% had at least one abnormality that had a match for a drug, and then they randomized either between, they either got a matched drug or they did not. Unfortunately, they did not see a difference in progression-free survival. And this was really concerning, and kind of had us all waving our hands in the air, like what does that mean? So the MD Anderson actually used retrospective data, but in a prospective fashion, and how do I mean that? So essentially, these are all patients that were on clinical trials at MD Anderson. They were all patients that had molecular testing, and then they followed what they got. So if you look along this table, there were 3,700 patients had a molecular analysis during this time, and do note, this was early, 2007 to 2013. Now I'm sure we have much higher numbers. Be that as may, two-thirds of them had some type of molecular alteration, and the majority of those were actionable, meaning there was some type of drug. Of those, 1,300 were actually treated, and 700 of them had a matched therapy. So a little over half of those patients got a matched therapy on trial based on availability. And you can see the most common aberrations were things as simple as ER, PR, or overexpression versus mutations in PIK3CA, BRAF, P10, and others. And I just want to note that about 20% of these were gynecologic cancers, and so somewhat relevant to our population. And here they saw that when those patients got matched therapy, it did improve survival. So they saw improvement in progression-free and overall survival, and again, in a very heavily pretreated group. And I will say that the regimens on this were often combo regimens. There were sometimes regimens with chemo, getting at that idea that maybe single agent targeting of a pathway may not be enough, and so doing those combination strategies might be helpful. So now I'm going to transition to some of the more current and some of the big successes, right? These are fun. These are fun slides to look at. So first was, of course, Pembrolizumab, which this very savvy audience is no doubt familiar with. Earlier, when we first saw these data back in 2015, we saw, of course, that mismatch-repair-deficient colorectal cancer got a ton of responses, had a ton of benefit from single agent Pembrolizumab. But also the non-colorectal cancers, which were predominantly endometrial cancer, had quite a bit of benefit. But there weren't that many of them. So they went on to do an expansion in endometrial cancer, which is in the table below, where they treated 49 patients and saw a 57 percent response rate for mismatch-repair-deficient endometrial cancer, which was, of course, very exciting. So not only did we get the kind of first tumor agnostic FDA indication from these results, but we also went on to get a specific indication in endometrial cancer. But what is kind of interesting is they also saw MSI in a lot of other more rare tumors, and I'm not going to jump onto Dr. Brown's talk because I know she's going to get into this, but, you know, if we test patients, we can find stuff. Even if the rate of MSI, say in ovarian cancer, is one to three percent, if you find that patient with MSI high, you're impacting her greatly by being able to know that single agent checkpoint inhibition is a good option for her. And so what they found when they looked at the unselected testing outside of, you can see the bulk of MSI was found in colorectal, endometrial, a little bit in prostate and esophago gastric, but then about 40 percent of the MSI high cases they had were in frequencies less, in cancers with frequencies less than two percent. So all those patients got on this trial, and a lot of them had benefit. Taking into a different direction, so larotrectinib, which acts on tract fusions. This is, I just would like to stay on this slide the whole time if I could. This waterfall plot was just, when they first showed this, when Dave Hyman presented this, it was so exciting to see, again, in patient populations that had really poor prognosis, multiple prior therapies, that this 80 percent response rate, which led to an FDA approval, again, in a tissue agnostic setting, and here we go, 65 percent. So we had 40 percent for the MSI high, 65 percent of the patients on this trial that had a tract fusion had frequencies less, or from tumors that have frequencies less than one percent. So it doesn't hurt to look, right? I mean, it's cost, we always have to talk about cost, but gosh, this patient population, what a benefit, you know, for tumors that don't have a lot of other standard of care options. These data were just presented at ASCO, and this is even getting more and more nitpicky, right? So we have, like, MSI high, and then we drill down to tract fusions. This is specific, a drug that is specific to a specific hotspot mutation in p53, and obviously p53 is widely aberrant in a number of solid tumors, including ovarian, uterine, our cancer types, but this is specifically drilling down to one hotspot. This is a key hot, mis-sense mutation, and it seems to destabilize p53, and you can see on the right the frequency of this particular hotspot mutation across a number of solid tumors, ovarian, one of the highest, okay, 2.9 percent. And here are the results. This was a phase one study of PC14586, which, as I mentioned, is a first-in-class p53 reactivator specific to this hotspot mutation, 32 percent response rate, median prior therapies of seven. So this is legit, right? This is a drug that works, and you'll see the majority of the patients actually have either disease reduction or disease stabilization, and overall it's very well-tolerated. The four things down is a patient with ovarian cancer. That's my patient. She's been on this therapy for over a year now, and she's tolerating it well, and she had gone through seven prior therapies without any benefit from anything. So this kind of very specific drill-down, understand the science and target it appropriately, these kind of trials are so important. So then we're also seeing kind of bigger baskets. We talked about the I-SPY trial where they can rotate different arms, and we see that with the Bokeh trial. For those of you that were in the ovary master class session yesterday saw Dr. Monk talk about that in rare tumors. We're trying to push the envelope and not get drilled down to one, two, three arms, one, two, three tumors. So you're seeing these combinations of basket and umbrella trials, and there's a number of different examples of this, but I'm just going to highlight two today. The first is the TAPR trial, which is run actually through ASCO. So it's the Targeted Agent and Profiling Utilization Registry. So it's a registry. The criteria are incredibly broad of who can be put on this. It's all based on physician discretion. There's nothing prescriptive about this trial. The physician simply says, I have a patient with an aberration, and I'm going to put her on something targeted towards it, and I'm going to collect her data for you. And it all gets put into this registry, and then as they see a certain proportion of patients with a given aberration, like say HER2, responding to a specific drug, they can report it. So it's really nice, and it really gets at what Dr. Myers today was talking about in the early panel. We're meeting the patients where they are. This lets any community physician, even someone that doesn't have the ability to run clinical trials, be able to participate in a trial that will impact getting these drugs out to patients. And so it's actually an IND-exempt trial, because these are drugs that the patient is getting, they're just prescribing in the clinic. But they do have very specific response criteria. So that's the one area they want to make sure that they can truly compare the data. And so they're very specific about that. But they really kind of get to where patients are. Now I think many of you may have participated in the NCI MATCH trial. Again, this is one of the earlier targeted umbrella and basket trials that included all solid tumors and lymphomas, and all of them were tested, and then they got binned into different arms based on drug availability. And we've started to see some of these trials get reported out. So here are some of the results. And similar to what we saw, say, in the SHIVA trial, you know, it hasn't knocked everyone's socks off entirely, but some of that is just based on what drug was tested. So you can see the first study of capanilacib, which is a PI3 kinase inhibitor, an IV PI3 kinase inhibitor. It was chosen by a PIK3CA mutation, and they saw response rates of 16%. This met the bar of interest, and this is going to be explored more, in fact, has gone into lymphomas and breast cancers and others. So that was great. The second trial of a fat nib that was in HER2-activating mutations, not quite as exciting, 2.7%. But, you know, it at least is moving the science forward, and again, getting patients onto trials and getting them access to drugs. I think, you know, the cruzotinib arm of the MATCH study was very interesting because it teased out either the patients that had an ALK mutation or ROS1 rearrangement and treated them with this cruzotinib, and you can see the difference, actually. Both had really reasonable response rates, but the ALK group, wow, 50% response rate. So really making an impact, and now we're seeing soon, it's sort of out in the public domain, so I didn't want to put the actual arms out yet, but there's the combo MATCH study, and I think we're all really very excited about this because our hesitance with single agent, we know there are mechanisms of resistance, and maybe that's why we're not seeing as much bang for our buck when we're targeting a specific mutation, so we're hoping that if we can use combination strategies that can overcome that resistance or prevent that resistance, we're going to see better outcomes for our patients. I'm not going to go through this in a ton of detail because Dr. Powell did an amazing job as always, but I think, you know, looking at platforms that are going to be able to not only provide us with the mutations and the protein immunohistochemistry to direct therapy, but also to potentially de-escalate therapy, to potentially add something new to kind of standard endometrial cancer treatment is very interesting. So I'm going to end on some of the more novel therapies that are coming down the pike that are going to require testing for our patients, right? So a lot of these trials that I'm going to talk about, the trial itself, the company itself did the testing, but what we'll be looking for are platforms that move quickly with the science and offer these same testing strategies to our patients so we can prescribe these important drugs. So I think, you know, we've heard a lot about antibody drug conjugates at this meeting. I'll go through the design of them very quickly. We have, of course, an antibody that's based on a target, so we find a target that we think is relevant in the cancer that we're looking at or the cancers, and that's linked to a very, very high-dose chemotherapy payload, something we could never give to our patients intravenously and allows for a more specific targeting of the cancer cells. We also see bystander effect where, as the payload is cleaved, that we can get impact on neighboring cells. Now this is a table that I stole from Isabel Ray-Cocard, and I've already been informed that it needs to be updated, so I'm going to get on that, y'all. But it does kind of show you some of the targets that we're exploring, including folate receptor alpha, NAPI2B, which are two I'm going to get into a little more detail because we have some really nice data in gynecologic malignancies, but also TROP2, which is highly overexpressed in endometrial and ovarian cancer, MUX16, which is CA125, obviously a very important marker, and others. And so we really do have a large portfolio coming down the pike. HER2 isn't even on here, and, you know, that, of course, is becoming more relevant for our patients with endometrial cancer. So let's talk a little bit, and you've seen these slides a couple times before, so I won't belabor the point, but hopefully I'll just complement some of the other talks you've seen. Mervitaximab, obviously, is an ADC-targeting folate receptor alpha, and I think that the development of this drug has been, you know, kind of, it's been a little bumpy, but it's somewhat aspirational because people, the investigators didn't give up, the industry partner didn't give up. So we saw really great early Phase I data for this folate receptor alpha ADC across ovarian cancer, and they went forward with Forward I, which was implatinum-resistant, one to three priors, with high or medium folate receptor alpha expression, and they were randomized to either Mervitaximab or investigator's choice chemotherapy, and the primary endpoint was progression-free survival, and it was not met. So those first two curves, you can see the intent to treat was negative, as well as the tease-out of folate receptor alpha high group. However, they realized that they had changed the way they were scoring, I don't know why it does that, changed the way they were scoring the folate receptor alpha, and when they went back to the PS2 scoring, they actually found that there was an improvement in progression-free survival with the Mervitaximab. So instead of just kind of throwing their hands up, they immediately transitioned and created two new trials. So SREA, which is a single ARN, and Mirasol, which is a randomized trial that will be the confirmatory study should the SREA trial data be accepted for FDA approval. So we've seen the SREA data, efficacy is good, 32% in those patients with folate receptor alpha expression, duration of response about seven months, and teasing out the subgroups, you saw benefit across lines of therapy, which is, of course, very relevant given our discussion around PARP lately, as well as if there was prior exposure to PARP. So really nice data that demonstrate that this is an active agent, and I think we're gonna hear in November if it's something that we're gonna be able to give to our patients. And then upafetamab rilzedotin, this is UPRI, I'm not gonna say that again. It's targeted, an ADC targeted to NAPI2B. Now NAPI2B is a sodium-dependent phosphate transporter, and it's very highly overexpressed in ovarian cancer. About two-thirds of ovarian cancers will have it. And it's not expressed very highly in normal cells, which makes it an ideal target. We don't know about, really much about overlap as yet between folate receptor alpha high, or folate receptor alpha expression and NAPI2B expression. So that will be something that we really need to understand as we're working on sequencing and potentially triaging patients to these agents, assuming we get our indications. So this was UPRI's phase one study. They did a phase one in ovary as well as lung cancer, and looking at the broad-based population, 67% of patients had target reduction. And if you look at the table below, you can see the NAPI2B high group had higher levels of response, so proving that this biomarker is appropriate. And interestingly, the dose mattered. So the lower dose seemed, the patients seemed to have better response, and whether that's due to being able to stay on the drug due to tolerability may be the case. But they've been able to really tease out the dose and tease out the right patient population, and they're moving forward with a number of trials that you heard about yesterday at the ovarian master class. And then the final thing I'm just going to touch on is adoptive cell therapies, because this is another place where having a platform that provides a lot of different data is really important. So adoptive cell therapies are really when we either collect T-cells from a patient and either grow or modify or adjust them in the laboratory and make them smarter, faster, stronger, better, and then put them back into the patient to fight your given disease. And we've seen this with TILs, where we actually take out tumor, take the lymphocytes out of the tumor, and then grow them and re-infuse them, or with CAR-T, where they're modified engineered receptors. But with a lot of these adoptive cell therapies, we've seen it get FDA approved across lymphomas and leukemias, but for solid tumors, it's been a little bit tricky. And there are a number of different steps to qualify. The first is a specific HLA subtype. The majority need some type of HLA-0201. And then we need a tissue. So you'll have your first pass of HLA, which you can get on some of the platforms, including the CARES platform. And then you send the tissue to see if they have the receptor. It's basically like giving the drug-sniffing dog a scent, so that it actually can find the cancer. So one agent that's kind of the farthest along, I would say, in ovarian cancer is ADP-A2M4-CD8, which is a CAR-T cell therapy that targets MAGE-A4. Very important in ovarian cancer, highly overexpressed, although we're still learning. We think it's somewhere around 30 to 40%, but it really depends on the patient population you're screening. And they presented their phase one data at ESMO this year, and saw really nice anti-tumor activity across multiple solid tumors that had that MAGE-A4 expression. You can see that it was fairly durable after T-cell infusion. And these are nice, right? This patient, they get treated once, and then they just get followed. It's a great thing. Once they get through that early side effects and toxicity, the patient's done, and she's just getting scanned. When they teased out the response rates based on tumor type, you can see for ovarian cancer, that which is on the left, it was quite impressive, where the majority of patients, again, had reduction in disease, including a complete responder in a heavily pretreated patient with ovarian cancer. So very exciting data, and this is moving forward into a phase three study. So bottom line, I think we've got a number of things that we can get from genomic testing. We can find these meaningful alterations, but we're still trying to understand which alterations are the most important, and how to best target them. So really being smart in how we develop our clinical trials and make sure we're being informative as we move forward is really the best option. And for you all, get the testing so that you can get the patients on the trials, or hopefully subsequently get FDA-indicated drugs. So thank you all so much for your attention. Dr. Wesson, that was fantastic, and inspiring that we're making real progress in precision medicine. Thank you. That took a lot of work to put all that together, so thank you. What do you think will be the next test that we're going to need on any of these reports that you get from the vendors? Yeah, I'm envisioning a panel of protein, you know, based on the different antibody drug conjugates that are in development, based on the different adoptive cell therapies that are in development. Because I think we're getting quite a deep dive from a mutation standpoint. I think that the transcriptomics is pretty in depth. But I think the protein, appropriately, has been minimal. There's a few things that we look at. There's a few things that are guiding our therapy. But as we start to see more and more of these positives, I envision a nice panel, an ovarian cancer panel, an endometrial cancer panel. But even beyond that, a solid tumor panel. Because as we saw, there's a number of different aberrations that may not be very common in a given cancer. But if you identify it for your patient, then you can give them an effective drug. Any questions from the audience? You would have to say, we hope there's not a companion diagnostic, which every one of these, it'll just drive us nuts. It's a really good point. Yeah, I'm hoping that we'll be able to incorporate it in the testing that we're already doing. Or even in our own laboratories, if we can, like we've done for PD-L1, for example. So one of the things that comes up on clinical trials is the requirement for re-biopsy and re-sending, re-sequencing, whatever we're looking for, looking for specific mutation. Or even with some of the IO to classify PD-L1 and so forth. Can you speak to the challenges of re-biopsy in the different cancers, just real quickly? Yeah, absolutely. I mean, I think, obviously, from a patient standpoint, having to go through the procedure and the cost and the discomfort and everything is one big thing. I think that the other issue is, how many times can you put somebody through those things? And this is coming from somebody who writes trials where we have baseline biopsy, on-treatment biopsy, progression. But those patients are heavily counseled, and I think when we're, and they're a special breed of patient that's going through that kind of a trial. What we need to do is move forward with a validation and a trust of, yes, if you're doing this test the same way as I'm doing this test, we don't need to send this in, we don't need to do additional testing. But I think it is on the trial development that you want to validate the test, you want to make sure you're using the right test. I would be remiss if I didn't at least mention, liquid biopsy obviously seems to be something that there's potential there, where we can, it's just a blood draw, just a blood draw, but it is a lot easier and a lot less invasive. And so if we can validate those things, I'm hoping most clinical trials that are doing this kind of re-biopsy are also getting the blood and the serum necessary to potentially correlate it to CT DNA so we can move away potentially from the invasive testing. Yeah, and it seems like the technologies for doing that are ahead of where we've been in GYN. It seems like in other areas there's been a much greater, yeah. Yeah, colorectal lung, there's definitely, at least from a prognostic standpoint, and now also they're changing treatments, they're starting to change treatments based on clearance of the circulating tumor DNA or not. And I think hopefully we'll have some more data in gynecologic cancers over the next few years. Well, thank you very much. That was a fantastic session there. Thank you. So our next speaker is Dr. Jubilee Brown, who hails from the Levine Cancer Institute at the Atrium Health, where she's Professor and Director of the Division of GYN Oncology. Dr. Brown's going to speak to us about the utility of molecular profiling in rare gynecologic tumors. Thank you. Great. Thank you so much. Thank you, Tom. And thank you to everyone who's here, actually, right now. I think it really speaks to each of you, not only for wanting to take the best care of all of your patients, but perhaps especially patients where we don't clearly always know what to do, and certainly that our rare tumor patients fit into that. So I'll be sharing with you today just some thoughts on the utility of molecular profiling in rare gynecologic tumors. So first, we'll talk a little bit about the spectrum of rare gynecologic tumors, and then I'm going to share with you about a case of a patient with clear cell carcinoma, and then some data on rare gynecologic tumors specific to this platform. We'll follow with a case on low-grade serous carcinoma, and then just a few summary words. So when we think about rare gynecologic tumors, the NCI has really helped us understand first how to pronounce rare cancer, but they've given us the definition. So what do we talk about when we're talking about rare tumors? Well, these are cancers that occur in fewer than 15 out of 100,000 people each year. And the EU has slightly different, even a more restrictive definition of less than six per 100,000 people each year. So where does that put us? Well, we in gynecologic cancers, not all of our larger buckets, like ovarian or uterine, are considered rare tumors, but really we have a lot of sub-buckets in gynecologic cancers. So when we think about uterine cancers, for example, it's not all endometrioid, it's serous, it's MMMT or carcinosarcoma, it's clear cells. And ovary cancer, when we think about the sub-buckets there, we think about our epithelial tumors certainly as being most common, but even within epithelial tumors, there are rare cancers. So whether it's low-grade, clear cell, or mucinous. So really, we're talking about a wide variety of tumors, and these present specific challenges. So it turns out that 35% of deaths occur from the rarest 20% of tumors. I mean, let that sink in. We have to do better for these patients. And what I'm going to share with you today is that this may be the path forward where we can do that. We have other challenges with studying rare tumors, and not just studying them, but knowing what to do with them. There's a limited number of patients over a wide geographic area, often with limited access to experienced oncologists. We lump similar tumors together that might not actually respond the same way. So clinically distinct tumors we often combine in order to meet accrual goals. There is limited funding. There is prolonged accrual. I look a whole lot different now than when I started the GOG-264, for example. And there are restrictions on international data sharing. Some of that is getting much, much better, and so we're able to do trials like the Bouquet trial now on an international basis, but that hasn't always been the case. And so when we look at the spectrum of rare gynecologic tumors, look at, for example, uterine corpus cancers. These corpus cancers affect over 65,000 women each year, and so as a whole, uterine corpus cancer is not a rare tumor. But when we break it down, you've heard Dr. Powell give a really excellent detailed dive into corpus cancers. And one of the things that we say is that the TCGA data are so exciting, and they are, but that TCGA data doesn't really apply to rare tumors. And when you break down in the appendix the data regarding the specific histology and grade of the 370 patients in this most heavily quoted article, it turns out only 53 patients are serous cancers. And so when we try and base larger guidelines on just a handful of patients, it becomes really problematic. You've seen this shared with you as well today from the memorial data. And the key that I take out of this when I look at it is that there's really no consistent characteristic even in rare uterine tumors. All of this background really just to say it's hard to study and treat rare tumors. We've tried from a conventional standpoint. In fact, between 1987 and 2011, there have been 33 randomized controlled trials on rare tumors, 66 single-arm trials in addition. And so we've had some success. There have been 45 products approved not just for GYN but across all of cancer for 68 rare cancer indications. And actually that's increased over the years. And that's great news for people with rare cancers. But if we look at really the outcomes of those trials in GYN, just uterine for example, look at the clear cell carcinomas. We have very dedicated researchers including Dr. Weston here. But they don't always turn into positive trials. And so we end up just fishing. We try. We try lots of different things that seem like they might be good. Is there a better way that we can study these tumors and guide treatment for patients? Well, so that idea turns into the promise of precision medicine to try and tailor treatment to each patient so that we can identify targets. We can identify which patient might benefit from immunotherapy or chemotherapy or certain investigational agents. And so this brings us to case one. So this is a 57-year-old patient of mine who I actually inherited upon my transition to Atrium Health. So this was a 57-year-old patient who had been diagnosed with a stage 1a clear cell carcinoma of the ovary. In 2013, she had open surgery. So you know I didn't do it. Total abdominal hysterectomy, bilateral salpingo oophorectomy, and staging. So a 1a clear cell, 57 years old. How many of you would give no-adjuvant therapy to this patient? A couple people. How many would give three cycles of paclitaxel and carboplatin to this patient? More people. How many would give six cycles of paclitaxel and carboplatin to this patient? Some more people, probably myself included. But now, how many would do either three or six cycles plus molecular profiling? I would vote that everybody should raise their hand. And this was before that. And so this patient got six cycles of paclitaxel and carboplatin. So she actually had a really tough time. She had anaphylaxis to paclitaxel and docetaxel. And so she ended up receiving gemcitabine and carboplatin for six cycles and was NED in July of 2013. So she was followed, and then essentially two years later, she developed a recurrence in her para-aortic lymph nodes. She had a laparoscopic resection and radiation to follow. So would you send genetic testing only? Nope, nobody. Would you send molecular profiling only? Anybody? One? Would you send both? Yes, most everybody. Because this is a patient who had ovarian cancer who didn't have germline sent initially. Would anybody send neither? No? Good. That's a win. So this patient got both germline and somatic testing. So she got panel testing. And so here's the thing. So she ended up having mutations in EGFR and PD-L1. And so she had progressed through a couple of other therapies. She ended up going on the TAPR trial that Dr. Weston mentioned. Because of her mutation, she was treated with cetuximab. Unfortunately, she did progress. Then she was re-challenged after desensitization with paclitaxel, and she got carboplatin, and she progressed. So what now? She got nivolumab. Why in the world did we do that? If we look back at I.O., immunotherapy in ovarian cancer, there hasn't been a lot of data, really, up to recent times. And back when this patient was undergoing consideration for therapy, she had gone through a lot of different options. So at the time, the best clinical response rate in this small trial of 20 valuable patients, the best clinical response rate was 15%. Clinical benefit rate was 45%, with a 10% durable response rate. And this trial found that PD-L1 did not correlate. But of the two complete responders, one was a clear cell carcinoma. So here's our patient. Remember, I'm going to flip back. Remember, she was positive for PD-L1, and she was a clear cell carcinoma. So we treated her with nivolumab, and this was the disease that she had. She had literally like a golf ball, maybe almost even a tennis ball in her super clav nodes, and widespread disease. Six months after, it's gone. So just really a nice little vignette there to really demonstrate and drive home the importance of comprehensive molecular profiling. Certainly that treatment is not something that we would have done just, you know, grabbing it out of the bag of all the options that we have. But this really allows treatment based on genomic drivers of tumorigenesis. And we see, just like you just saw, dramatic responses in treatment refractory disease. Because this broad panel testing measures levels of genes or gene products, and actually turns into a biomarker of response, both prediction and prognosis. So KRAS was actually nice enough to share data on rare ovarian tumors. And so specifically just, we presented this at SGO two years ago. When we look at the data on clear cell carcinomas, what I want you to recognize is that there are a significant number of actionable mutations in the clear cell carcinomas that have been profiled. A lot of them, right? So you've got arrowed 1A mutations, and PIK3CA mutations, P53, KRAS, P10. Okay, so what I also want you to see is that it's not in everybody, right? And so how are you going to know, out of a room full of patients with clear cell carcinoma, how are you going to know how to treat them? Well, unless you profile them, you're not. In fact, that whole story holds true for all rare tumors, not even just ovarian tumors. And so if you sort of combine all these data, you look at the minimum percent of patients with targetable mutations, and whether you look at stromal tumors, where there are 10 percent of patients who have targetable mutations, or germ cell patients who have 20 percent of targeted mutations, even mucinous carcinomas, 62 percent. Like literally, if you go down the board, not just for ovary, but also for uterine sarcomas, for neuroendocrine GYN cancers, there are targetable mutations in rare tumors. But if you don't look, you will not find them. And you simply can't generalize. So you can't generalize even to, you know, the histologic subtypes that we have. All of these patients are different. And so we presented this, as I mentioned, in 2020. This really supports molecular profiling to identify potential targets, and supports, as we move forward, a platform trial strategy to study rare gynecologic cancers. And, you know, we have that now with the bouquet. It really sort of follows what you, Shannon, talked about with the I-SPY trial, where we have arms that can just plug in. So hopefully, we'll be able to sort of plug and play our way through evaluation of these different options for patients. It's based on a much stronger strategy than histology alone. So when we look at biomarkers of immunotherapy response in solid tumors, I think this also kind of drives home the point that when we're talking about immunotherapy, you know, we tend to say, oh, I'm looking at TMB, or I'm looking at PD-L1. But what this really highlights is that it's not the same, right? We really should be assessing multiple markers of immunotherapy sensitivity in order to maximize identification of patients who might benefit from this therapy. And I think that's really important, just another sort of important concept, so that if we have a rare tumor that we're looking for, are they MSI-high or MSI-stable? That's not enough. That's not enough. You need to have more information to appropriately direct your patients. And so this kind of brings up the comprehensive molecular profiling and just how it works. And so I wanted to just take a minute to kind of go through the boots on the ground, like the daily approach to how this works. So when you have a patient who you want to sample, you take a tissue sample. And this, basically, it's not hard. You know, this is something that the pathologist can simply gather for you from paraffin-embedded tissue. That sample can be sent to CARIS, and then the biomarker testing happens. So CARIS analyzes that tumor tissue, and we'll talk a little bit more about their multi-omic assays to identify biomarkers. We mentioned proteins, and that's sort of the wave of the future, but we're actually there. So we look at DNA, we look at RNA, we look at proteins. And even more, this goes into a big informatics analysis, and that's something that's very, very unique. So there are a lot of sort of big data that get compiled into the report that you get back, because that tumor profile is then generated. It's not complex. It's something that, you know, I share with my patients, and it's pretty easy format to read and review and interpret. And patients aren't confused when they read it, because it's just, it's color-coded in terms of what works. So this is the report that you get. And so based on the mutation analysis, the, it'll be color-coded. You see green if there's a potential benefit, and so again, it's easy to go through this in a short period of time, as we see probably too many patients in a short time period. So when we get this comprehensive molecular profiling, it's the combination of whole exome sequencing of DNA, which is important because that looks at mutations, it looks at indels, and it looks at copy number alterations. And it's 22,000 full gene coverage, so it's really whole exome coverage panel. The RNA as well, you get whole transcriptome sequencing, looking at fusions, looking at variants. And, you know, historically that hasn't been too important for us in GYN, but, you know, I think we see the rapid change that happens. So we are going to need this information as we move forward, especially in our rare tumors, which as you see, the mutations that present are really just across the board. And then protein, you know, immunochemistry, immunohistochemistry is also incredibly important for us to have. Dr. Herzog mentioned microdissection, and so this really speaks to tumor enrichment of the specimen that's performed by tissue microdissection. So you don't, basically you don't have an inadvertently negative sample that's analyzed. And that's really important. You don't want to get the information on benign tissue. You want the specific tumor that you're looking for. And so in this way, specific tumor cells are identified for testing, and so you don't get dilution of the tissue section. This is especially important in our settings of limited tissue. So, you know, the question is, oh my gosh, how much do I have to send? Well, not a lot. You can actually a lot of times get core biopsies that you can send if we want to re-biopsy or do, you know, sequential biopsies on our patients. Another just thing, it says whole exam sequencing, but that is spell correct, sorry. That's a whole exome sequencing for TMB and MSI, and so I think this is really important in terms of the sensitivity, specificity, and positive and negative predictive values. These are sort of best in class. So this brings up case number two. So here we have a 55-year-old female. So back in 2015, she had a stage 3c low-grade serous carcinoma in the ovary, and I actually inherited her as well in a different location even in Charlotte. And she had had an exploratory laparotomy, lysis of adhesions. Apparently it was just terrible. She had no mental biopsy. Tumor was unresectable. So she had paclitaxel and carboplatin for six cycles, and of course, that's not surprising, right? Low-grade serous carcinomas, they can be very fibrotic. So she had KRAS testing sent. She was ER positive, PR positive, and KRAS positive. So how would you treat her in this setting? She's had her interval tumor-reductive surgery. She has residual tumor in the omentum. She's already gotten six cycles of paclitaxel and carboplatin to try and shrink her down. So how would you treat her? Would you do surveillance? How about hormonal therapy? Yeah, a lot of takers for hormonal therapy. How about Bevacizumab or paclitaxel carboplatin more? All right, so it sounds like hormonal therapy has it, and that's what she had. And you know what? It worked great for five years. So she had five years with just stable disease on letrozole, but she recurred then with a vengeance. She had progression with massive ascites. She ended up coming in and getting paracenteses every two weeks. She couldn't work anymore. I mean, this is a young, active, healthy female, except that, I mean, you know, she got huge amounts of ascites. And her CT scan then showed a nine-by-six-centimeter mass. She was in a lot of pain, couldn't function. What would you do now? Surveillance? Bevacizumab, anybody? Lots of ascites. How about paclitaxel and carboplatin again? How about Tremetinib? Yeah, that is my answer. And so here's what we have. So here was her pre-Tremetinib scan. You can see, again, she's getting paracenteses every two weeks. She's got a mass. She's got tons of ascites. And literally, shortly thereafter, this is her most recent scan, but it did not take her long. It took her about two months to dry up. She has no evidence of disease. She is doing beautifully asymptomatic, and actually, you know, now that we know more about Tremetinib and how to sort of tweak it, we've minimized her side effects, and she's tolerating it extremely well. She has a little bit of a rash, but that's about it. So I think to summarize the take-home messages here, there are targetable mutations in rare tumors. So you see that. And you cannot generalize. So every tumor is absolutely different, and personalized therapy is just essential. These mutations guide our therapy, and they are clinically meaningful. So it's, you know, we've come a long way from just theorizing about what might be useful. And this is useful, and this translates, as you see in these two little patient vignettes, into improved patient outcomes. So with that, I'd like to thank you and open it up for any questions or thoughts or commentary, not just about rare tumors, but really about anything. Thank you so much. That was fantastic. My last slide didn't make it in, because I sent it probably too late. But one of the things I wanted to talk about was that the genomic forum that just happened in San Diego the last couple of days, and it was headlined by former President Barack Obama as the keynote speaker. Just amazing amounts of information coming out. But one of the things I thought was really interesting, you know, we talk about costs in doing sequencing, and we've also been interested in this whole bioinformatics of how we store all this data. It's just amazing the new things that are coming out. For example, the NovaSeq 6000 that most people use, they're coming out with a NovaSeq X and a NovaSeq X Plus. And first of all, it takes 44 hours to run a full run on the older one, and you get about 6 terabytes of data. The new ones that are going to ship in 2023 are going to be able to be done in 24 hours or less and get 16 terabytes of data, terabytes. Cost is about the same as the old ones, a little more, but the Plus I think is $1.25 million. The other ones are a little under a million. But the cost of sequencing one gigabase has gone down significantly, so much so that we're probably looking at potentially a $200 human genome sequencing, which is just phenomenal. Now, there'll be other costs with that because a big part of the cost is the reagents and so forth, but even there, they're moving to thermal stable solutions where we may be able to get away from dry ice and all that styrofoam and packaging for environmental purposes and everything's going to be great. But just to show you the tremendous technology improvements that we continue to see in this area. Like I said, it's going to be a very different world as we move forward. I know we're at the top of the hour, but any burning questions from the audience? Jubilee, just one question I had for you. I think you showed that nice Venn diagram, and maybe we can just close on that, is that when you look at, especially in the I.O. space, I think we're really still struggling, right? Disease by disease. You alluded to the PD-L1 really doesn't have a big place in endometrial cancer in terms of being predictive. Other areas we see that's quite the opposite. Tumor mutational burden, MSI high. How do we come up with some type of weighted process for understanding? You saw some of those were hot spots where they were all, patients' tumors had all three, for example, and then others just two, and then some just one, and what's the breakdown? Because it seems to be different between all our gynecologic malignancies. Yeah, no, that's true. This is working. That's okay. I'll be loud. I don't know that we have any one specific answer yet to that, but I think what's clear is that one test is not enough. So when there's a panel to be able to look at that, and that's what we have here, when there's a panel to be able to look at that, it gives us more information, and so we don't overlook somebody who might benefit from immunotherapy. Great. All right. Well, we're at the allotted time, so we'll close here. I want to really thank the panel, Drs. Brown, Weston, and Powell. Thank you for really well-prepared talks, and just a great job. So thank you, and I want to thank the audience for hanging in there on a Saturday afternoon in New York City on a sort of nice day. Thank you.

molecular profiling

gynecologic cancers

treatment decisions

precision medicine

predictive markers

molecular testing

endometrial cancer

clinical trials

targeted therapy

rare tumors

panel testing

personalized therapy

genomic drivers