So, hello. I'm Adrian Suarez, Director of Surgical Pathology at The Ohio State University Wexner Medical Center in Columbus, Ohio in the United States. I'm the current chair of the IGCS Education Pathology Workgroup. We thank you all for being with us here today. For today's webinar we have a real treat. I'm joined by Dr. Charlene Bosse, who is an opinion leader in gynecological pathology and endometrial carcinoma in particular. Dr. Bosse, please introduce yourself. Can you hear me? Yes, I can. All right. Well, hello everybody. I want to thank the organization for inviting me to talk. My name is Charlene Bosse and I work at the Leiden University Medical Center as a pathologist with expertise in gynecological pathology. I'm looking forward to the next hour or so. Thank you for joining us. I would like to congratulate you on your many accomplishments and contributions to the field and awards, including the prestigious Young Investigator Award by the Dutch Cancer Society, the Helga Saabesen Award by ESGO, the Jeremy Jazz Prize for Research Excellence by the Journal of Pathology, and more recently, the United States and Canadian Academy of Pathologists Ramsey Cotran Young Investigator Award. We really appreciate you joining us here today and sharing your insight with the IGCS global community. Before we get started, I would like to mention a few housekeeping items. Please know that a recording of this webinar will be available on the IGCS Education 360 Learning Portal within one business day. We encourage you to submit your questions via the Q&A feature at the bottom of your screen, and we will do our best to address as many of those questions as possible. Dr. Bosse, I will now hand the webinar over to you. Please share your screen. Thank you very much, Adrian. I will start sharing my screen, my presentation. And so give me a second to get my technically started. And maybe you can confirm that all looks good. Looks great. Fantastic. And I'm also going to activate my pointer, so I have a pointer during my presentation. Thank you all for joining this webinar. I'm going to entertain you hopefully for the next 45 minutes or so and then have some time for questions. I am reaching out from you from the Netherlands, from my office at the Leiden University Medical Center. It's a shame that we can't be together in a physical room, but maybe this will happen in the near future. It's a very timely topic to talk about endometrial cancer, and in particular the newly developed molecular classification that has really resulted in a paradigm shift, I think. I want to thank the IJCS once more for inviting me, and I'm looking forward to this next hour. And it's also timely to talk about endometrial cancer because endometrial cancer is on the rise, and that's not just locally, but it's a global rise we're looking at. Very recently, published in the Lancet, Emma Crosby from Manchester published a very nice review article in which a lot of the very recent data on incidence of endometrial cancer is described, and as you can see there's a clear rise in the diagnosis of endometrial cancer worldwide, with higher income countries going up, but also low and middle income countries are starting to rise. And in more detail, Dr. Crosby looked at the low and middle income countries and saw that there's actually a market increase, although There's a market increase in those countries that are not usually mapped as those with the highest incidences, and they report that there is clearly a direct link to the obesity pandemic that is going to overwhelm us with new diagnosis of endometrial cancer. So what we're looking at is a major healthcare burden worldwide and therefore it's very timely to discuss this topic and what we can do, especially if you also take in the US numbers into account in this paper from a few years ago, where there was a projection made on incidences towards 2030, and as you can see in the dotted lines, endometrial cancer will surpass colorectal cancer by the year of 2030. And again, mainly due is the current hypothesis due to obesity. Okay, I'm a pathologist, and I feel centered around the management of endometrial cancer at Molecular Tumor Board, because of course at the preoperative biopsy I give the diagnosis, which results in the cornerstone of treatment, which is surgical removal of the uterus. And then subsequently, I again feel myself centered in the discussion about adjuvant treatment, because there are different modality options that we can choose from. We can, after the removal of the uterus, decide not to do anything, follow up, we can give radiotherapy, we can give chemotherapy or a combination of both, and we can also consider novel targeted treatment these days. And it's been, for a long time, been a decision that is based on a risk model, which the pathologist actually provides critical elements. And these are stage, of course, how deep in the myomedium is the tumor evolving, is it evolving in the endocervix involved, are the adnexus involved, but also tumor characteristics like histological subtype, FICO grade, grading from one to three, and the presence and extent of lymph vascular space invasion. And together, this composite of aspects make up an individual patient's risk. And this risk, often divided into low, intermediate, or high risk, subsequently impacts the management decision. And this is how endometrial cancer has been treated for decades. However, already for some time, people are discussing that this is suboptimal, and is mainly suboptimal because of the people like me, we have problems with this, because there is a high inter-observer variability, actually for all aspects, not just histological subtype, but also grade. But also grading, the assessment of lymph vascular space invasion, the assessment of stage is highly variable, and there's a large body of evidence that this is the case for many different cohorts, and in many different countries, it's been reported. So that's a really, actually poor basis for management decisions. In addition, this has actually, this whole model lacks a biological basis, which is problematic if you want to move forward into more personalized and directed treatments. So about 10 years or so ago, we thought of the idea that maybe a molecular approach would be part of a solution as it is more robust, there's less inter-observer variability, and it would provide a biological basis. And in fact, this has been incorporated in the WHO classification, this is kind of like the Bible for us pathologists. In 2016, it was clearly histologically driven, where we classified endometrial cancers based on either them being type 1 endometrioid, or one of the type 2 cancers, of which the dominant one, of course, is uterine serous. And now in 2020, and we are already in 2023, of course, we moved away from this classification into a classification that is fully based on molecular factors. And this, of course, is a major paradigm shift, and tries to fulfill this problem, or tries to provide a solution towards the problem of inter-observer variability. Now, I'm going to start with my first poll I may have to introduce. This is a question to you guys out there listening to this webinar, could you please participate? There are five answers to my question, who are you? And this can help me with the rest of my talk. So are you a non-believer, let's say, with or without resources, doesn't matter? And would you feel more comfortable, would you identify yourself with the fact that you're not really believing that the molecular classification adds any substantial evidence to my clinical management decisions, and therefore, HND is maybe old school, but sufficient? Or are you number two, you are a believer, but you don't have any resources? Or are you maybe number three, a believer with limited infrastructure resources? Or four, believer with a modern level laboratory, but I don't get reimbursed? Or are you number five, you're a believer, and you completely have implemented the WTO classification? Please vote. And now I'm supposed to wait 30 seconds before Ashley can provide us the results. So I'm curious if this is going to work. This is the first of five polls that I have, I think, or four in my presentation. So I can't see, Ashley, if people are voting. But if you feel comfortable that you have enough answers, then please share. Ah, here we go. Very nice. So I got a lot. My job is going to be quite easy because the majority of you are believers. So that's nice. But then there is the problem that I hear a lot. 63% has limited infrastructure or resources. So hopefully, after this talk, I can provide you with some possible solutions for the obstacles that you might encounter. I also see that there are some people that actually are already implementing the molecular classification, 15%. So that's also very nice to hear back. All right, let's continue. My slides. Yes. So back to the molecular classification and the basis of it. In 2013, the TCGA published their landmark paper in Nature, where I think close to 300 endometrial cancers were completely comprehensively molecularly classified and characterized. And I think this is, of course, a very busy figure, number one, with lots of information in there. But what is an important take home is that every single endometrial cancer is molecularly very diverse, very different. And what they have done, of course, is they have been able to sort of structure this chaos into four groups for us. And these four groups are on, they are clustered based on tumor mutational burden and copy number alterations. And what we see is on the far left hand side for me is the the the pol E or ultra mutated tumors with an ultra high mutational burden and with the characteristic mutation in DNA polymerase epsilon, which we call pol E. Then there are the highly mutated microsatellite unstable tumors and on the far right hand side are those that have very little mutations actually often limited to p53 and a lot of copy number alterations, which is referred to as copy number high in the TCGA. Now very early on, we already looked at this problem and decided that fairly practical tests can be performed and used as surrogate markers to sort of parallel this classification by the TCGA, you can do targeted sequencing for pol E, you can do immunochemistry for the MMR proteins to identify microsatellite unstable tumors and you can do p53 to identify the copy number high tumors and that what leaves you are the tumors that have been called no specific molecular profile or NSMP, which sort of parallels the copy number low from the TCGA. So this is what I refer to as the molecular classification doing these tests. And was this is this prognostically relevant? Well, in the TCGA wasn't really made for it, but you saw quite striking stratification, particularly, of course, the flatline of the blue pol E cases was striking, but this was in the TCGA, I believe, only 17 patients. So very quickly thereafter, we performed the surrogate marker approach using two big randomized control trials, PORTEC1 and 2, more than 800 cases and found very, for early stage cancers, a really striking stratification. And to stress this and to make this even more exciting when we look at only grade three endometrioid and endometrial carcinomas, and this was an international collaboration I did back in 2018, you also see this stratification. And here is always the moment where I say that when you treat a patient based on the histological diagnosis, grade three endometriot and endometrial cancer, you're actually treating one of these five, and they are quite different, one of these four, I should say, and they are, of course, quite different in their prognostic behavior, if you have a closer look. And by the way, it was, I think, 15% that was pol E mutant in these grade three tumors. Now, a lot of publications have followed and some also have done unselected retrospective cohorts, pretty much validating and revalidating these findings of the molecular classification. So really, the molecular classification has significant, adds significant prognostic information. And then very recently, we also showed this in the PORTEC3 randomized control trial, in which patients were randomized for chemotherapy with radiation or radiation alone, only high risk based on classic features. And very strikingly, again, the molecular classification stratifies into four prognostic risk groups. Sometimes people say, well, but maybe the p53 group wasn't staged appropriately, because in PORTEC3, staging, full staging by lymphadenectomy wasn't necessary. But if we look at a recent publication by Alicia Leon, where she looked at cases that had full lymphadenectomy, you see, still see poor outcome for the p53 low stage, even after a full lymphadenectomy. And then an often asked question is how does this apply for non endometrioid morphology, a little bit less numbers, smaller numbers in the literature on this, of course, because of its rarity, but at least what we can say is that poly mutations exist in every single cohort, also in non endometrioid cohorts, and they reproducibly show an excellent outcome. And for uterine carcinose sarcoma, I saw at the USCAP in New Orleans this year that Brooke Howard from Stanford has a large cohort of more than 100 uterine carcinose sarcomas with nice stratification on the molecular classification. So yes, it does perform in non endometrioid morphologies. Just as a side note, also in endometrioid type ovarian cancer, the molecular classification has prognostic relevance. Now, just a few words about the hierarchy of the testing. As you know, the TCGA had this hierarchy, and this was taken over by the WHO. This is from the WHO, and here you see that the testing algorithm actually has a hierarchy. It starts with poly testing, subsequent MMR testing, and then P53 testing ending into one of these four molecular classes. Now, this hierarchy question has been discussed in the literature. This is because you can have something we call multiple classifiers. So a scenario where you have either a poly or MMR deficient tumor that also has a P53 mutation. So you got the good and the bad in one case, and how to deal with that. Now, Alicia Leal tackled this together with a large international collaboration, and this work was highly cited and received the Jeremy Jess Prize from the Journal of Pathology. And in this work, she collected a large series of these so-called multiple classifiers. It's not easy to identify them because they're not that common. And what she showed is a couple of striking findings. Number one, this is just about the MMR-D-P53 combination, this slide. You can see quite commonly in these cases, more than 60%, that the P53 stain, when you do the immunochemistry, shows subclonality. What do I mean with that? You have a part of the tumor being mutant and a part of the tumor being wild type. And that was actually quite a striking finding. And if she did hierarchical clustering of the copy number alterations, you see that the double classifier in brown clusters with the individual single MMR-D and not with the single P53 abnormal endometrial carcinomas. And then the clinical outcome of the multiple classifiers, so the combination of MMR-D and P53 in this case, if you compare that to a reference group of P53 singles, you clearly see that they follow more the line of what you are used to for single MMR-D. And this strongly suggests, of course, that the MMR-D-P53 combination should be classified, should not be classified as P53 abnormal. And we can actually overlap almost this story for the combination polymutants and P53. Very striking, of course, these findings where the additional mutation in p53 doesn't drive the behavior of the tumor. It still has an excellent clinical outcome. So this is the hierarchy. The WHO took it over. The ESCO, ESTRO, ECP guidelines have connected treatment management decisions. So we need to start, and we need to start doing it. Which brings me to the next question, polytesting. Are you already testing every single endometrial cancer is answer number one. Are you testing a selection right now? Or are you not testing at all? And I'm going to give, again, 30 seconds to think about this easier question. Are you reflex testing number one? Are you having a selection? Or are you not testing for poly at all? Because the WHO actually tells you you should. The guidelines, the European guidelines, at least. All right, Ashley, you have results? Yes. All right. So although I have a group of believers in my audience, 60% is not testing at all. And there's only a small group, 4%, who's testing every endometrial cancer. And then there's, of course, the middle group, who probably has the possibility to do testing, but does it for a selection. All right. So we'll talk about that a little bit. I need to get control back. Yes. All right. So polytesting. When you do polytesting, you have to do it correctly, of course. And one aspect that I need to talk about, of course, is when you do find a variant in poly, you have to make sure that it's a true pathogenic variant. A lot of the variants actually fall also in the MMRD cluster and are actually not driving the tumor. So when is a variant driving, or is it considered a true poly mutant endometrial cancer? Well, it has to be inside the exonuclease domain, because that's where we see the ultramutator phenotype arising. It needs to be associated with signature 10 and associated with ultra high tumor mutational burden. Now, I know you're not doing whole genome sequencing, so you wouldn't know these things necessarily in all your cases, which is why the work of Alicia is very nice, because she worked out for all the available data at the time, what variants would qualify as a poly mutant. And she found indeed that poly mutants have a certain signature, and this resulted in a list of recurring variants. These are 11 variants that all qualify as a poly mutant endometrial cancer, because they have shown to result in an ultramutator phenotype as well as give the appropriate signature. Now, by far the majority of the cases you will identify are in the top five of this list, and then after that it's becoming exceedingly rare to identify other cases. So, but it's important to remember that if you're sequencing that not necessarily all poly variants qualify, so if they're not in this list, you have to be cautious. Cheaper alternatives, because I'm assuming the obstacle for full implementation might be costs, I put in this slide. There's alternatives for NGS, because of course if you only have 11 hotspots, you can do hotspot type of work. There's different commercially available machines that can run poly testing for you. We also just recently developed a very simple qPCR test, fully validated with high sensitivity and specificity, just got published in JCO Global. Everybody can read the methods. It's very cheap. It has a turnaround time of two hours, and it pretty much drops the cost of poly testing to the level of immunochemistry, and you only require a simple PCR machine for it. Now, I'm a pathologist, so I always wanted to just have a look at how do these tumors actually look, so that's what I did back in the day together with an old PhD of mine, Inge van Gogh, and we looked at 150 cases, and I was blinded for the microscope, with 50-ish cases polymutants in the mix with others. So what's very striking for these tumors is that they're filled with these little dots, as you can see in these nice H&E images. Those are lymphocytes. There's tons of tumor infiltrating lymphocytes, as well as in the tumor, as outside the tumor, with these nice tertiary lymphoid structures here in panel B. So that's very striking. There's other striking aspects, such as scattered tumor giant cells, which may cause a pathologist to call it even a serous cancer, and what we concluded in this paper is that the majority is high-grade endometrioid. There's lots and lots of tumor infiltrating lymphocytes, but we also had to conclude that sometimes other histological subtypes are preferred, like carcinosarcoma, serous carcinoma, etc., etc. We did note often a subclonal p53 stain and also subclonal, sometimes MMRD stain. So if you do want to select your tumors for polytesting because you have limited resources, a possible approach would be to limit your testing to grade 3 endometrioids or grade 3 endometrioids with lots and lots of lymphocytes, or grade 3 endometrioids with lots of lymphocytes and subclonal p53 expression. And with every step, the chance of identifying a poly is increasing. Of course, we did some additional, let's say, research work. This is me with Inge Vogel and David Church back in Oxford, back in the day, where we looked into this problem on how is it actually possible that these tumors have such a good outcome. Of course, we followed up on these histological findings, and as many of you right now know, a lot of these tumors have high influx of CD8 positive T cells. This is what this work shows, both intratumoral as well as in the surrounding stroma. And what was interesting at the time is that the RNA expression data of the TCGA showed upregulation, actually, of immune suppressive molecules, and very significantly so. And Brooke Howitt in Stanford showed this also on the protein level with high expression of PD-L1 and PD-L1. And if we say that, we automatically think of checkpoint inhibition. And that's what people have done. There's a couple of case reports of advanced-stage polyimmuted endometrial cancers showing really excellent responses to checkpoint inhibition, not very surprisingly. But the majority of your polyimmutant endometrial cancers will not be advanced, but will be early states. So what do we do in terms of adjuvant treatment? Well, let's go back to PORTIC3. We had two treatment arms, chemotherapy plus, radiation plus chemo, CTRT, or radiation alone. Well, it's very obvious that we shouldn't be giving chemotherapy to polyimmutant endometrial cancer patients. So drop the chemo immediately. But then, of course, we're still left with radiation. Are these patients or these tumors super sensitive to radiation? The answer is no. We looked at this in vitro in cell lines where we made polyimmutant cancer cell lines. In this graph, you can see that the mutant cell lines, they are not more sensitive to radiation than the blue wild-type control. And we also kind of knew that if you look at the PORTIC1 no additional treatment arm, patients that didn't get any adjuvant treatment and were polymutants had an excellent outcome. So no evidence for hypersensitivity to radiation. And this was later also confirmed by Alicia in the what we call Danish cohort, which also didn't get adjuvant treatment. So this is all the basis for the RAINBOW studies you may have heard of. I will shortly discuss some of them. But if you want to know more details on the RAINBOW program, please find the publication in the International Journal or go to the recently published podcast with some of the PIs, which will give you all the answers on the questions you might have on the RAINBOW program. But of course, it has, for example, the polymutant blue trial in it, which is open already in Canada and international activation follows this year. And it is going to de-escalate adjuvant treatment prospectively in polymutant endometrial cancers. But of course, the community needs to test in order to recruit these patients. Yeah, you might be thinking that's all nice, but I'm still having the trouble that it is so costly. Well, the money saved by treatment de-escalation can easily be covered by the costs of poly testing for endometrial cancer. And this is also not just my statement, but this is also comes out of the cost effectiveness studies that are being published now. So let's move to MMRD-MSI, mostly endometrioids. Again, mutation load is high, not as high as the polys. And also, this tumor group has upregulation of, for example, PD-L1. MMR is being tested by either immunohistochemistry or by microsatellite instability testing, such as it was or it started back in the day with the test markers calling the tumor either MSI low or high. Right now, the immunohistochemistry has somewhat of is recommended because it can identify subclones and it can direct the clinical geneticist's trajectory. Which brings me to the next poll. MMR testing in your institutes. Number one, reflex testing on all. Number two, reflex testing on a selection, for example, only under the age of 70 or only a subset of tumors. Only when requested, which is slightly different. So only when a specific clinician is asked for it. Or are you number four, MMR testing is never performed on endometrial cancer in my institutes. It's a good moment for me to have a break. It's nice to have a poll. I can't see Ashley, but in the background, Ashley is looking at the results. And there they are coming. Yep. Okay. So here's a bit different than the poll. There's a lot more MMR testing. I kind of expected that 62%. But still only. So I'm surprised to see that 26% answers number three. So only when requested. So that's interesting for me. And then there is a small group that never tests MMR. Okay. Interesting. Let's see if I can get the control back again. Yep. So I'm a big fan of MMR testing already for quite some time. As a young pathologist, I was in the audience of my MSK colleague, Dr. Robert Soslow. And when he started talking about MMR, he said it was a pathologist chance to save lives. Well, that might be a little bit, you know, strong, but at the same time, he has a point. It's very relevant, clinically relevant to identify patient at higher risk of Lynch. Of course, if they have Lynch, you want to prevent them from getting second cancers like colorectal cancer. Oftentimes, endometrial cancer is the first cancer. And there can be preventive measures for family members. And actually all the international societies recommend MMR testing independent of any variable. 3% of your total population has Lynch. Within the MMRD population, 10% has Lynch. Applying an age cutoff is possible, but it would miss 50% of your Lynch syndrome families if you apply the 70-year cutoff. Personally, as a pathologist, I find it difficult to say I'm going to test the 69-year-olds, but not the 71-year-olds. And subsequent methylation testing, if you do that, will identify a group that has within the MMRD group a slightly worse prognosis, although it was not statistically significant in our cohort. A bit about MMR testing. So it's for immunochemistry stains. You can have a couple of unusual patterns. You can have subclonal loss of MLH1 and PMS2, such as in this figure. That's mostly due to promoter hypermethylation. You can have triple loss, where you have MLH1, PMS2 loss, and subsequent subclone of MSH6 loss. Nothing crazy about that. There's a microsatellite area in MSH6. And then you can have individual loss of MSH6 and or PMS2, which is often caused by pathogenic variants. This staining, by the way, when performed in a good laboratory with a skilled pathologist, has a really, is very accurate and has very low inter-observer disagreements. It's interesting that you can have this subclonality. We've seen that this MMR, when it's subclonally lost, you also see in this area, microsatellite instability. The other area is stable, and it's probably due to tumor progression. And then, of course, there is the possibility to identify MMR-deficient gland foci. That's usually in the context of lynch, where normal glands show loss of expression, and it's an indicator of Lynch syndrome. Now, when it comes to making things affordable, let's have a good look at how this actually works. The MMR proteins, they form dimers, MSH2 with 6 and MLH1 with PMS2, which has the subsequent result, you can never have individual loss of MLH1 or individual loss of MSH2. This is our cohort of 1,200 cases that we looked at back in the day. And this means there is a possibility to do a more cost-effective approach, where you take a two-antibody approach. You do the stains, MSH6 and PMS2. When they are proficient, 70% of the cases, you are done. When they are deficient, you continue on with the rest of the testing algorithm. This is a very sensitive approach. We did a meta-analysis, and it's a very safe and appropriate approach, and it reduces your cost by 35%. All right, after we've identified the MMRs, how are we going to actually treat them? Well, it looks like we shouldn't be giving them chemotherapy either. If you look at the treatment arms of PORTIC3, there's no benefit of chemotherapy in this group. So we have to look at alternatives, and the alternative is around the corner because a lot of new data, and this is already a slightly outdated slide because the publications are out now. These are for advanced and metastatic MMR deficient endometrial cancer, and we see really exciting results using checkpoint inhibition, and of course, and these are here tabulated, and of course, this is also going to move into the adjuvant setting, in this case, in the rainbow MMRD green trial, which is open in the Netherlands, and other centers are going to be activated in 2023, which brings me to the copy number low or no specific molecular profile tumors. This is sort of the mixed bag of endometrioid low-grade tumors with a lot of Wnt and PRT kinase activation. Not a lot of non-endometriotes in there, but they do occur, and Lisa Vermeij, one of my PhDs, just recently published a nice paper on the landscape of NSMP endometrial cancers. This is only the high-risk NSMPs, not the low risks, and there you can see she made the proposition that you can make a distinction between those that are ER positive, which is by far the majority, and those that are ER negative here on the right, and you see a slight difference in their profile. The ER positives are your, let's say, standard endometriot type endometrial cancers, but the ER negatives are a little bit uncommon and different. They have a little bit of a different molecular profile, and under the microscope they tend to also look a little different. They're high-grade and have non-endometrioid morphologies. This is, again, that subdivision. I think it's about five percent, but it could also be a little lower in a more population-based cohort, and for example, the ER negative NSMPs, you will find the mesonephric-like carcinomas and some of the clear cell carcinomas fall in this category. And does it matter? Yes, it does matter. Actually, the ER negative NSMPs, they have a very bad outcome, very much comparable to p53 abnormal endometrial cancers, but the majority of the NSMPs are hormone receptor positive, and therefore it might be an interesting target, and this is just a proposal that maybe we are moving towards a molecular classification algorithm that starts to also include ER. Not only our group, but some other groups have confirmed this finding, so the field is moving forward, let's say. How are we going to treat the NSMPs? The benefit of chemotherapy in this group is not so convincing. There might be some in product three, it wasn't super clear, and therefore the NSMP orange trial is being designed. This is a non-inferiority trial in which we are trying to see if we can take away the chemotherapy in this group and maybe replace it by hormone treatment, given their hormone receptor expression. We're ending, of course, with the most aggressive group, the p53 abnormal group, and we're also ending with my last poll question. P53 testing, are you testing reflex? Are you doing it on selected cases? Or are you never testing for p53 in endometrial cancer in your practice? I'm very curious whether this is going to give a little bit of a different picture than the MMRs. The international societies aren't as clear in their instructions, maybe. It gives me some time to have a look at the time. We're good in time, looking good. All right, Ashley, you have results? Yes. Oh, well, I'm happily surprised. 60% is reflex testing. That's quite a lot, actually. 29% on selected cases, and only 12% never tests P53 on an immutable cancer, which is also quite a striking percentage, actually. All right, interesting. All right, so when you do P53 testing, you know that there's different outcomes possible. You can have your scattered nuclear expression, which is wild-type, and then there's several possible mutant-type expressions. There is, of course, the classic nuclear overexpression, which can also be subclonal. And then on the far right, we have two examples of more rare situations where you either have complete loss of expression of P53, or you have a cytoplasmic accumulation of P53, and those two are also indicative of an underlying mutation. We know from the work of, among others, Navina Singh, that this model actually is very well reproducible, very accurate to identify mutations, so there's no real need for mutational analyses, and therefore, P53 is an excellent marker. It doesn't matter in your reporting whether it's a null mutant or an overexpressing P53 mutant. We do not have any indication that that's the case based on the work we did, so there's overlapping curves here. I mentioned earlier that P53 can have a subclonal expression with both a mutant and a wild-type area. This can explain, for example, discordancies when you do NGS with a punch biopsy or of a micro-dissected area, and these mutant areas can also be multifocal, and when you see this, it's a very high chance you're actually dealing with an MMR-deficient or a polymutant tumor, which I tend to call together one of the mutators, and the P53 here is probably a secondary hit. Now, this is a very large and interesting paper from MSK, who has put together a really large cohort, almost 240 P53 mutant endometrial carcinomas and their molecular profile to look at potential targets. There's all kinds of possibilities here, but the one thing that I always like to highlight when I show this slide is the bottom, because I'm a pathologist, and the thing that people tend to forget is only 50% of the P53 mutant endometrial cancer are actually called serous, which leaves the other 50% as non-serous P53 abnormal, which includes grade one endometrioid endometrial cancer, and we also see that, this is just one example from a recent study we did, that it is possible to identify P53 mutants low-grade endometrioid-type endometrial carcinomas. How are we gonna treat P53 abnormal endometrial cancers? Well, here, the PORTEC3 data are very clear that these patients truly benefit from the addition of chemotherapy to the radiation, as you can see in the Kaplan-Meier curve, and that alone would be an important reason to do the molecular classification, because it identifies patients that can benefit from chemotherapy. Now, the outcome is still poor, so we should probably look at alternative targets. You may have already seen work on HER2 in endometrial cancer. Indeed, there is about 20, 25% HER2 amplification in the P53 abnormal subgroup. It's almost mutually exclusive with the other groups, so there's no need for HER2 testing outside P53, and there's trials underway, particularly in America, to study the potential benefit of HER2 targeting in this group. Another, let's say, possibility is to look at ovarian cancer, high-grade serous, and we see there a lot of benefit or good results with the use of PARP inhibition, and we think that's due to the occurrence of HR deficiency, homologous recombination deficiency, and we have studies showing that HRD is also occurring in the context of P53 abnormal endometrial cancer, which, of course, has been the rationale to do the P53 abnormal RET trial within the RAINBOW program, which is opening in France and will be opening international also this year. And the complete RAINBOW study program is also sponsored by AstraZeneca. It's an academic trial, but AstraZeneca is providing the drug. So I'm coming to an end. I want to first sort of summarize for you the cost-saving possibility solutions to potential obstacles that you encounter. I assumed that the obstacles that you encounter are mostly due to limited resources, so therefore I'm highlighting these. I gave you an option for a cheap hotspot QPCR method to identify polymutants. Bringing the polytesting costs down to an immunochemical cost only requires a PCR machine and doesn't need any molecular personnel. I shared with you the two-antibody approach that can save costs for MMR testing. P53 immunochemistry, if you look at the diagnostic algorithm, can be limited to MMR-proficient endometrial carcinomas because the second hit in the context of MMR is not driving your molecular class. And I shared with you several references to which it shows that treatment de-escalation pretty much covers the complete implementation, so therefore the whole approach is cost-effective. So the solution really is the molecular classification. It does add costs and turnaround time. It remains, in some scenarios, a problem. Some other people described to me that the obstacle of the molecular classification is that it only looks at the tumor and not in its microenvironment. So when we look at the future, maybe even other solutions need to be considered. One of particular interest would be to integrate artificial intelligence or deep learning into the field. And this is what we are currently doing and working on. This is Sarah Vermont, an excellent PhD of mine, who has recently published in the Lancet work in which one H&E goes into the computer, digitalized H&E, of course, and the model predicts the molecular class. And this was done by training and validating a large cohort of a population of more than thousands endometrial cancer, one slide per patient, and then tested on a completely unseen test set, in this case, PORTEC3, to look at model performance. And we can see that the average area under the curve is actually quite high. This is for a complex model that needs to predict one of four, very high performance. And if you look at the individual predictions, particularly P53 is pretty accurate already. And this is only the first version of this model. And we are already working on and seeing better results in subsequent versions. And if you would only use this one H&E based, image based predictions, you can already get quite good stratification. So the future in that sense is bright because this will reduce the costs enormously because you won't need any molecular testing at all anymore. So I think, I hope I have discussed and provided some solutions to possible obstacles that you encounter in your practice. And I think we can all really start to perform and implement a molecular classification in our institutes. I'm happy to take questions. These are my acknowledgements. Lots and lots of people work together with me internationally who I think, and of course, also all the patients that participate in the trials that we use, the PORTEC trials, and for the future, the RAINBOW trials. Thank you. Now, I don't know if I- Thank you so much, Chavi, for your wonderful presentation. We have multiple questions here. So let's see how we do. So first one here is, what is the value of doing POLE testing in FIGO grade one endometriate carcinoma with intact MMR, P53 wild type, stage T1A? Well, that's an excellent question. So if you go from the assumption that, so in that context, you've done the MMR stents and P53, so that's taken care of, and you're left with your morphological observation that this is an endometrioid grade one tumor superficial, so super low stage. In that scenario, POLE testing might not add much. I always, when I think about this question, I always think, you know, if it were one of my friends, or if it was my aunt, I still think there's a difference between low risk and no risk. So if assuming that POLE testing drops in price, I think it would be quite nice to have that information. So that will be one part of the answer. The other element, of course, is that a lot of people are moving towards testing pre-operatively, and then you don't know the information about stage at that time point. So then, yeah, you may have already started the testing. And then, well, I guess the person who's asking this question is right. It's a scenario where it's not super relevant for the clinical decision-making. And then I also kind of like to refer to things like breast cancer, where we do ER, PR, HER2, and KEY67 on every single case. And it's also not clinically relevant for every single case, but we're still doing it because it's sometimes simpler to organize when you have large volumes. Thank you. So our next question, for initial surgical management, should B53 abnormal low-grade endometrioid carcinoma be treated as a high-grade tumor? That's a good question. So this refers to the work I only shortly showed, and I didn't show the whole story. It's shared at USCEP by Lisa Vermeyer. So just shortly, so what we did there is we looked at the cohort of these cases, low-grade endometrioid, which six expert pathologists agreed were low-grade endometrioid, but B53 mutant. And we looked at the clinical, and so we collected, I think it were about 50 of those cases, and we collected them and looked at clinical outcome. And they showed a significant worse outcome than what you would expect from a low-grade endometrioid low stage. So basically, I can only answer the prognostic question. They don't follow the prognostic line of a B53 wild-type grade one endometrioid. Whether they will also be, whether they should then also be treated with chemotherapy is difficult to prove at this stage, but I think this is the way we are currently thinking, yes. Great, thank you. Have you seen loss of the four MMR markers simultaneously? Yeah, it happens. Not common, I must say, but I've seen it once or twice. Definitely the three is very common. So what I mentioned, so MLH1P-MS2 plus MSH6, because there's a microsatellite hit, a frameshift hit in MSH6, and then subsequent loss of MSH2. I have seen it once or twice in my career, must say. Yeah. Perfect. So here's a longer question. Based on ESGO-ESGTRO guidelines for the management of patients with endometrial carcinoma stage 1a B53 abnormal without myometrial inmation, considering immediate risk group, the question is, if there is a stage 1a B53 abnormal with myometrial invasion, but less than 50%, do you still classify in this group or do you consider high risk? Yeah, so at the time the guidelines were written, there was a little bit of a lack of data on this particular group. It actually refers back to the previous question. So it's all about, you know, everybody's struggling with this, okay? So they are low stage. They are not deeply invasive. They don't have LVSI, but they are B53, how to deal with that, right? And it's uncommon and therefore difficult to study. That's with rare entities, always the case. The most recent evidence that we do have seems to suggest we should classify them probably as high risk and therefore also treat them as high risk. So I'm expecting actually a move in that direction in the next guideline. Regarding ER immunohistochemistry, what's the cutoff value that we should use and what intensity? Nice questions. Very, very good questions. So ER, as mentioned, is sort of, is like back to the future, right? So it's a new kit for the molecular classification. At the time the publications came out from ASCO, ASTRO, as well as from the WHO, the role of ER wasn't worked out. And just recently, we and others have shown that the role of ER might add a layer of information, particularly in NSP. Now, we had applied 10% as a cutoff, and we don't look at intensity at all. This is also because I don't think that we are very good at intensity scoring. So it was also from a pragmatic point of view that we decided not to look at intensity, but just at percentage of expression. And to be honest, we looked at the different cutoffs and we saw the strongest sort of stratification with 10%, which is why we went for the 10% in our publication. I know there's another publication that proposes 1%. So I think the jury is still out. I think a lot of our cases under the 10% were actually plain negative. So I think in the coming years, we will probably, with our international community, find out what is the optimum, but it will probably be somewhere in that it's either gonna be one, five, or 10%. That's what it's gonna be. Right now, I use 10%. So, and for the time being, then weak staining counts. Yeah. If it's more than 10%. Correct. All right. So I think we have time for one more question. Do you think that there should be a specific recommendation for the subclonal loss of MMR? All cases with this pattern should be sent for MSI? Question mark. Yeah, it's a difficult one. So subclonal, there's a broad range of subclonal. You can have, let's say, a very tiny bit of 5% of your tumor and then the chances that it comes back as MSI high is small. And in fact, it's probably an indication of a mutator phenotype. So what you should be testing is poly in that setting, actually. If it's 80 or 90% loss, then you might be looking at tumor progression and that might simply be an MMR deficient endometrial cancer that's at the initial, initiation was MMR proficient, but the deficient clone is overgrowing the tumor. Right. So let's see if we can squeeze one last question here. So do you recommend poly mutation testing for all endometrial cancer at your place? Are you testing all? Yes, we are. For many of the arguments I've mentioned during my talk, there's a difference between low and no risk. The molecular classification without poly is not complete. You're gonna have to do it for every single case that you're looking at. So you have to do it for every single case you're gonna have to do it for every single case that is P53 mutant anyway, because it makes such a difference in management. We need it for recruiting in the RAINBOW trial. So there are so many reasons why that it's too difficult for me to think with every case, do I need it or not? In most cases I do. So therefore I just do all. Okay. And do you test biopsy or hysterectomy specimens? Yeah. I must say I'm moving more and more towards biopsy specimens in my practice. That's mostly because of course DNA is of the best quality, but also because then I have the data of the result already at an early stage in the treatment decision plan. So during tumor board after hysterectomy, then it's already known and it's kind of nice. So there's no delay, but that depends of course on your turnaround time for your molecular lab. All right. So I think that's all the time we have for today. That's too bad. Thank you so much. I wish there was more time. Thank you so much. It was wonderful. Thank you everybody who has attended. The recording for today's session will be available in the IGCS 360 learning portal within a day. And next slide, right Ashley. And this is just a reminder that registration and housing is now available for IGCS 2023 in Seoul, South Korea to attend either in person or on demand. We wish you continued health and safety and thank you so much.

molecular classification

endometrial cancer

treatment decisions

genetic markers

POLE mutations

MMR status

P53 abnormalities

prognostic significance

patient management