Okay, good afternoon, everyone. Welcome to your first postprandial session of the afternoon. I'm Daniel Leung from Baylor College of Medicine, Texas Children's Hospital, and it's my honor to co-chair the session with Dr. Jessica Zuckman-Rossi from University of Paris, Descartes. Welcome to our session today, entitled Pediatric Liver Cancers from Mechanisms to Treatment. This will be the first of two slated sessions devoted to the study of pediatric liver cancers. We really hope that this is going to be a state-of-the-art overview. It has been jointly proposed by the Liver Cancer and Pediatric Liver Disorder SIG, and we really hope to highlight the multiple pathways that lead to hepatic carcinogenesis, which hopefully will inform histological subtyping, molecular classification, as well as resistance to current therapies. And yes, there is a second part of the program, which will start, unfortunately, an hour later. We'll give you an hour break, and then the second session will resume at 3.30. And the second session will have a focus on improving upon our current therapy. So I hope you'll be able to join us. What's unique to this session is that we are going to defer questions for all the speakers to the end of the session. We will have a devoted panel discussion lasting 20 to 30 minutes, so please do come with questions. The open mics are available there. And I will now turn it over to our distinguished moderators, Drs. Paul Monga from UPMC and Drs. Kalyani Patel from Baylor College of Medicine, who will introduce our speakers for this afternoon. Thanks for joining us. All right. Thank you and welcome. I'm Paul Monga from University of Pittsburgh Medical Center. I'm in Department of Pathology and Medicine and run the liver center there. So it's a pleasure. We're going to, as Daniel pointed out, we're going to hold off on any questions until the open forum, which is going to be at the end of these four talks. So our first speaker is Kirk Vangenstein, who moved about a year ago from UPenn to Mayo Clinic and is interested in CRISPR screening to identify the genes and drivers of liver growth. And he's going to talk to us today about germline mutations and risk of liver tumors. Kirk. Okay. I'd like to thank the organizers for the invitation to give this lecture on germline mutations and risk of liver tumors. And as a disclaimer, I'm an adult gastroenterologist, but I specialize in cancer genetics and I'll do my best to cover this topic. These are my disclosures. The learning objective is to describe the germline or inherited associations for hepatoblastoma and hepatocellular carcinoma, which are the predominant or the main primary liver cancers of the liver in pediatric populations. The main cancer is hepatoblastoma, and it accounts for about two-thirds of primary liver cancers in pediatric populations. And most of the rest of the incidence is covered by hepatocellular carcinoma, but there are more rare tumor types that we may hear about later. But my talk will focus on the genetic associations for these two cancers. And I just wanted to quickly note that the incidence for hepatoblastoma has been rising over the last couple of decades for unclear reasons. It's been proposed to possibly be due to increased incidence of low birth weight babies, which is an association for sporadic hepatoblastoma. So here I'm listing the main germline genetic associations that have been linked to hepatoblastoma, but most of hepatoblastoma is thought to be sporadic, that there is no germline association linked to it. In the main association, the most prevalent one is with familial adenomatous polyposis syndrome, or FAP, which is linked to mutations in the APC gene. And hepatoblastoma has also been described with Beckwith-Wiedemann syndrome, glycogen storage disease, X-linked mutations in GPC3, trisomy of chromosome 18, and on a case report level with mutations in PKHD1, P53, and BRCA2. In perhaps the largest series of patients with hepatoblastoma that have been examined, this is actually a series done by Dr. Zuckman-Rossi's group among French patients with hepatoblastoma, you can see the germline mutations that were identified. They included mutations in APC, GPC3, BRCA1 and 2, and Axin1. And I want to note that these mutations are cancer predisposition genes. So in contrast to hepatocellular carcinoma, most of the genes are thought to be linked to cancer drivers rather than liver injury mechanisms. I'll discuss a little bit more about FAP, which is due to its autosomal dominant condition due to mutations in APC. Only about 1 percent of patients with FAP develop hepatoblastoma, but on the flip side, among patients with hepatoblastoma, up to 14 percent are found to have FAP. And this can be the first manifestation of FAP for patients and sometimes for their whole families. So it's important to detect these mutations because there are important implications there for colon cancer risk later in life. APC germline and beta-catenin somatic mutations are mutually exclusive, and that makes sense because they both drive the same genetic pathway. And beta-catenin mutations are common in sporadic hepatoblastoma, and I think Dr. Monga will be talking about that. And finally, hepatoblastomas in association with FAP tend to have increased immune responses to cisplatin. I think Dr. Zuckman-Rossi will be talking about that. And they have a good overall prognosis. So it's recommended that all patients with hepatoblastoma be tested for APC variants, and in children with FAP, that they should undergo screening for hepatoblastoma with an ultrasound of the liver. It's not entirely clear why only a small fraction of patients with FAP ever develop hepatoblastoma and why it only presents in childhood, but there does seem to be a relative absence of mutations in this mutation cluster region indicated here as compared to classical FAP. Moving on to hepatocellular carcinoma, as I mentioned, the germline associations in pediatric patients generally tend to be in genes that cause liver injury and cirrhosis. So this includes alageal, glycogen storage disease, progressive familial intrahepaticolastasis, and hereditary tyrosinemia. I'll also discuss that family history of HCC has been linked to increased odds of developing HCC. So in this series of nine patients, again, from Dr. Zuckman-Rossi's group, a majority of patients were found to have pathogenic germline variants in genes linked to liver injury phenotypes. Alageal syndrome is an autosomal dominant condition due to germline variants in JAG1 and NOTCH2, and it has a number of syndromic features, including a triangular-shaped face. There can be variable penetrants for specific organs, phenotypes in the liver. The liver disease occurs from a paucity of bile ducts, which can lead to cirrhosis. And the risk for HCC occurs mostly in patients with cirrhosis, but exceptions have been described. And so screening with ultrasound is recommended for all patients with alageal. Hereditary tyrosinemia type 1 has a very strong association with HCC. This is an autosomal recessive condition due to bialelic loss of function of the FAH gene, which leads to a defect in tyrosine catabolism and accumulation of fumarole acetoacetate in hepatocytes, which is toxic. There's a treatment that blocks an upstream enzyme and prevents formation of the toxin. In the absence of treatment, up to 75 percent of patients are found to have HCC. But if noticinone is started early in life, for instance, if it's picked up on newborn screening, the outcomes are generally quite good as long as noticinone is taken for life. Screening, however, is recommended with ultrasound and AFP. Next, with progressive familial intrahepatic cholestasis, or PFIC, there's an association with HCC, especially for PFIC type 2 due to mutations in ABCB11, which has been reported to have an incidence rate of as high as 15 percent. The mechanism is thought to be through bile salts exposure to hepatocytes, leading to injury and genomic alterations. And these types of HCC has also been described with other PFIC variants. For these patients, screening is recommended with ultrasound and AFP. Moving on to epidemiological data showing that a family has shown that a family history of HCC increases the odds of developing HCC by more than twofold, and this appears to be independent of the injury or the presence of hepatitis B or C in the family. So this suggests that there are genetic factors that can be inherited that increase the risk of HCC, even though a lot of the risk can be attributed to these acquired factors like hepatitis infection. So I led a study to examine the germline mutations in cancer-associated genes for more than 200 adults with HCC that we saw at a tertiary medical center. And we were surprised to see that more than 11 percent of patients with HCC had pathogenic or likely pathogenic variants in these cancer-associated genes, including significant associations compared to non-cancer populations for FANCA and BRP1 genes. And I highlighted here genes with especially high penetrance for a variety of cancers, including breast cancer and colon cancer, which would be important to know because patients can be cured of their HCC by liver transplant, but they may still be at increased risk for these other cancers. And we found similar results at a commercial testing lab where patients were referred for genetic testing due to their HCC. And even though there may be referral bias here, about 13 percent of patients were positive for one of these pathogenic germline variants, including these high penetrance genes that are really important to know for the patient and their family. In addition, it may be useful to know these germline mutations as they can potentially impact on treatment. This is a case report of a patient who had HCC, and their HCC progressed on serafinib. The patient was found to have a FANCA germline pathogenic variant, and based on the result from other cancer types showing that defects in FANCA and homologous recombination leads to increased responsiveness to PARP inhibitors, this patient was started on PARP inhibitor and cisplatin, and they had a dramatic response of a tumor biomarker and disease stability for 12 months on this treatment. So it indicates that germline genetic testing can impact on therapy. And then, finally, I'm part of a study group that's examining genomic data from the Million Veterans Program for common genetic variants, SNPs, that may be linked to HCC risk. So they compared almost 4,000 patients with HCC to matched controls that have established cirrhosis, and they found that there are SNPs in PMPLA-3, the TERT locus, and near the MYC locus that were linked to elevated risk for HCC. So it's possible that a polygenic risk score can help to risk stratify both adults and children with cirrhosis. This study will be presented on Monday morning by David Kaplan. So in conclusion, I think it's important to consider the genetic basis of cancer risk in patients with liver cancer, and to screen for liver tumors in patients with genetic risk. Now, there are no existing single gene panels that are available to test for these genes in patients that are comprehensive. I think that's a focus of study for my lab and for many other labs that hopefully will lead to more information in the future. So with that, I'd like to acknowledge my research team, my collaborators, my new employer, which I actually only started there in July, but it feels like at least a year so far, and my funding. So thank you very much. Thank you, Kirk. That was great. We'll hold off the questions until the end of this session. So our next speaker is Dr. Kalyani Patel from Baylor College of Medicine. She's a pediatric pathologist and is going to talk to us about, and I'm waiting for her to open it, histological subtypes of pediatric liver tumors. Thank you, Dr. Mangal. So no financial disclosures. In the interest of time, I'll just head directly into the talk. There's a lot to cover. Pediatric hepatocellular tumors can be sort of roughly divided into benign, intermediate, and malignant categories with entities within each of them. So starting with focal nodular hyperplasia, this is the second most common benign liver tumor in children after hemangioma, a wide age range. But typically, we see them in seven to nine, sometimes up to 12 years old, a good deal of female predisposition here in some cities as high as eight is to one in females to males. Pathogenesis is still poorly understood, but vascular abnormalities, drugs, and environmental factors have been implicated. And risk factors remain risk factors in some of our patient populations are those with hematopoietic stem cell transplantation, history of prior chemotherapy, non-vascular abnormalities, and female gender. So here are some studies. This is from North America with 12 patients, 79 patients from China, and 50 patients from Europe. As you can see, the most common is still 73% occurring in healthy children, a good amount of female predisposition between hematopoietic stem cell transplant and chemotherapy together, I think, kind of equals to vascular disorders. And what was interesting, this study from Texas Children's, is we found a good number in children less than five years. And that sometimes can be a difficult diagnosis because 75% of tumors in this age group are malignant. So it can be very challenging to assign a benign diagnosis for a liver lesion in this age group. Also, FNH in this age group can many times be atypical by radiology. They can be larger in size. They can show mild AFP elevations and growth during puberty in older children, but factors that make diagnosis, a clear diagnosis, difficult. Classically, they say these are five very typical imaging features that one should look for, similar attenuation and ecogenicity to the background liver. The lesion would be overall quite homogenous. There is a strong arterial phase enhancement with no washout. A central scar when present definitely helps a lot in absence of a capsule. And so when you have atypical lesions, especially in smaller children, they typically get biopsied or resected. In adults, they are typically considered do not touch lesions, but in children, we often see them biopsied or resected. So here is a very classic appearance up here as a gross imaging of a focal nodular hyperplasia with a nice central scar. The periphery of the liver shows proliferation of very bland, monotonous hepatocytes, and the central scar is usually richly vascular. The vascularity is actually quite typical in many ways. So you can have angiometas appearance. Many times, we see thick vessels with this myxoid fibroinimal hyperplasia or thickening, cavernous dilatation of the vessels. Sometimes we see thick arteries. Usually there are more venous channels, but sometimes we can see thick arteries. We have also had lesions with intralesional intravascular thrombi. They can be ossified, calcified. And so imaging can pick up these changes. And sometimes they can be really helpful. If you evaluate the peripheral, the adjacent liver carefully, you can see changes of obliterative portal venopathy. So here is a portal tract in the adjacent liver showing a nice bile duct, a nice artery, but no clear portal vein. And so that helps, again, in the diagnosis. Central scar can show abundant bile ducts, highlighted by cytokeratin 7 and 19. This is a helpful feature when you have a small biopsy to look at, because the needle can go through only periphery, only central scar, or a little bit of both. And so important to remember that the picture can vary depending on the sampling. Glutamine synthetase stain really helps in the diagnosis in a good number of cases, not all, but this is described as classic geographic map-like pattern in the lesion. And the central scar typically will be deficient of glutamine synthetase. So again, to remember, you know, again, sampling really affects diagnosis. Periphery of the liver showing normal glutamine synthetase staining in the perivanular pattern. And so coming to some atypical features in FNH. So this is an example of FNH where fibrosis was just maybe a little too much than what one would expect, and maybe not limited to the center, but also extending elsewhere. And because of this fibrotic and sclerotic sort of areas, the hepatocytes almost look a little laminated and was misdiagnosed as fibrolamellar HCC. So again, some, you know, something to think of. I think one of the most common pitfalls I have seen in diagnosis is actually FNH mimicking cirrhosis. And Dr. Leung knows we actually just had a case two weeks ago where the patient was diagnosed as cirrhosis twice. You know, biopsy performed at one institution, then referred to another institution, called cirrhosis in both places, and we were evaluating the patient for transplant and it was an FNH. So definitely to keep that in mind. FNH, when the central scar is not the best or, you know, very minimal, they can mimic an adenoma, and especially in lesions where, as you can see, the glutamine synthetase does not show a classic geographic map-like pattern. And there is a, I would say this is not a settled issue. People who, there are sort of two schools of thought where people would say, okay, any amount of central scar is sufficient, and then there is a controversy where if you have a minimal central scar and an atypical glutamine synthetase pattern, should you just call them adenomas? And to that, we had previously what was called angiometas FNH is now called telangiectatic adenoma. I think what is important is the clonality studies in FNH, and most have, you know, very small, or a small percentage show clonal cells in FNH lesions. More importantly, there are no somatic gene mutations in the beta-catenine TP53APC or the HNF1-alpha identified to date in large series. Isolated, some examples are there, but not in large series. So this coming to hepatic adenomas, most of our knowledge on hepatic adenomas is actually coming from the adult data. You know, pediatric data on adenomas is limited. These are the standard four categories that we know of, and, you know, we have Dr. Zakman Rossi here. They expanded the proposed and expanded classification of the hepatocellular adenomas with additional mutations in these genes. So here's a good, I would say, a reasonably, you know, large five institutions over 11 states, 31 patients on pediatric-only hepatocellular adenomas. That showed about third were prepubescent. This was very interesting, and we have seen this in our center as well, where if you're post-pubescent, there is a good amount of female dominance, but in prepubescent, we actually see male-female sort of equal, sometimes even more males than females. The third half were syndromic, again, non-syndromic. There is a great deal of female preponderance, but in the syndromic cohort, male-female is usually the same, and these were the subtypes in this series, most common being inflammatory. So coming to some examples, this was a transplant of 18-year-old female transitioning to male, morbidly obese, was on oral contraceptives for five years, and now on testosterone therapy, and had multiple hepatic masses, was underwent transplantation. As you can see, the adenoma has some periodic areas, very bland hepatocellular proliferation here. CRP is a good marker. It's called C-reactive protein, nicely positive in the adenoma, in the lesion, and LFABP, which is the liver fatty acid binding protein, another helpful stain for adenomas, is retained and similar in expression to the adjacent liver. So based on these findings, you know, we would call this an inflammatory subtype. Another example here of a 14-year-old, again, obese girl with multiple hepatic masses underwent transplantation for this, and as you can see, very bland proliferation of hepatocytes within the masses. Here are some areas of steatosis or evacuated hepatocytes, and LFABP is lost compared to the previous example where it was retained. Here it is lost, and that, you know, in our institution, we do molecular testing on pretty much all of our adenomas, but especially if you have loss of LFABP, I think these are good cases to do either targeted or, you know, a panel testing for HNF1-alpha gene, and we found mutation in this gene, calling this an HNF1-alpha subtype. Here is another example of an adenoma developing in a syndromic patient. This is a five-year-old boy who had congestive hepatic fibrosis due to a DCDC2 mutation, neonatal sclerosing cholangitis, incidental masses on the explanted liver. Now, interestingly here, LFABP is absent, and so, you know, you would call this HNF1 likely HNF1-alpha inactivated, and that was also seen genetically as well. In pediatrics, we have also encountered some hybrid lesions where you have biopsies showing FNH-like and adenoma-like areas, you know, and so they need very careful evaluation. And coming to another example here where beta-catenin is negative, there is loss of LFABP, so with the loss of LFABP and beta-catenin negative, one would think this is HNF1-alpha mutated, but no, this was a beta-cat mutated adenoma. And so, it also shows the fact that beta-catenin immunostain is not a reliable marker for beta-cat mutation. Here is an adenoma with very focal cytologic ATP and beta-catenin nuclear reactivity. We call this beta-catenin mutated not an adenoma, but a diagnosis of atypical hepatocellular neoplasm, which is a newly proposed entity basically for adenomas that have focal cytologic ATPR, focal small cell change, so basically features that are insufficient for the diagnosis of HCC, but beyond what one would expect in an adenoma. I'm going to quickly go through HCCs, but non-fibrolamellar HCCs are much more common in pediatric than fibrolamellar. This is an example of a three-year-old boy with tyrosinemia where, you know, the liver showed multiple nodules. And as you can see, this nodule here is quite different from these nodules showing significant cytologic ATPR, nuclear crowding, hyperchromasia. You can see the difference in reticulin and CD34 and glyphecan-3 between these nodules, so the upper nodule here would be called a large regenerative nodule, but the lower two nodules are basically HCCs. And this is to emphasize even internal heterogeneity. Some nodules can be very strongly positive for glyphecan-3, and some would be totally absent, you know, negative for glyphecan-3, important to consider if the child is being considered for a CAR T cell trial for glyphecan-3. Fibrolamellar HCCs are less common in children, but we do see them. CK7 and CD68 are good markers, and typically in our experience, glyphecan-3 is negative, so typically they do not, you know, satisfy the eligibility criteria for the GPC-3 trials. The PRKCA3 event, fusion event, helps in diagnosis, and there's no consensus on the outcome differences between fibrolamellar and non-fibrolamellar HCCs. Hepatoblastomas, we have epithelial, mesenchymal, mixed epithelial, mesenchymal, teratoid, and HCNNOS. Fetal hepatoblastomas are typically larger cells, bland, monotonous, vacuolated cytoplasm. You can have mitotically active fetal, microtubular fetal. Embryonal areas typically will show extramedullary hematopoiesis. As you can see, there are differences in beta-CAT and glyphecan-3 between fetal and embryonal, so that helps in diagnosis and should be carefully looked for. The significance of identifying fetal and embryonal is illustrated in this case very beautifully. This was a large right lobe mass with a very tiny segment for nodule, 8 millimeters nodule here. And the large mass was just pure fetal, cytologically bland, no ATP, mitotically inactive. And so if the patient only had this, just resection would be curative, no need for chemotherapy. But because of the presence of this small nodule here with embryonal component, 3 to 5 mitosis, chemotherapy is necessary. Mesenchymal differentiation can be seen as osseous, adipocytic, cartilaginous, teratoid as in the presence of melanin, neuroglial differentiation, squamous differentiation, goblet cells. We don't have a large series or good data to say that presence of mesenchymal or teratoid components affect prognosis. We have settled this issue of small cell undifferentiated a while ago. This was being identified and, you know, there were even studies that quantified the amount of small cell undifferentiated component. Now we know it does not adversely affect the outcome in hepatoblastomas. Post-therapy changes, pleomorphism can be seen especially in the fetal nodules in hepatoblastomas. And this should be very carefully assessed because this should not bring in the different diagnosis of HCC based on pleomorphism only. We have a newly proposed entity called hepatocellular neoplasm NOS where are these pleomorphic tumors sort of in between hepatoblastomas and HCCs. I won't go into many details but a good amount of studies are underway for histologic and biologic recertification. So, you know, basically they are hepatoblastomas with cytology KTPR tend to occur in older children, tend to be chemo-resistant, tend to have poor outcomes, and tend to have third promoter mutations, chromosome 1 Q gain. And so to summarize, tissue diagnosis is crucial, biopsies if possible should be taken from lesion transition and non-lesion, histologic diagnosis of FNH needs good correlation, age can influence adenomas, age and syndromes can influence adenomas in children, atypical pleomorphic neoplasm must be considered for those adenomas with atypical features, and histologic subtyping is necessary for appropriate management. Here are my acknowledgements, funding from the SIP-RIT, Dr. Feingold and Dr. Lopestre-Adam, my local mentors here. Many thanks to the SIG committee for the invitation including Daniel Leung. And we will take questions in the end. Our next speaker here is Dr. Sardarshan Paul Monga. He is the director of Pittsburgh Liver Research Center. He has worked for over the last 20 years elucidating the role of Wnt signaling in liver pathophysiology, and he'll be speaking on the role of beta-catenin in the development of hepatoblastomas. How do I get out of here? I just realized there's no laptop here. Let's click on this here, forget about that, all right. So I have disclosures, some of the work that I'm going to discuss today was done in collaboration with L. nylon. So I think we have heard already from our first two speakers about the complexities of these liver tumors in pediatric space, and I just want to focus on one particular tumor, hepatoblastoma. And again, just to reiterate, it's the most common pediatric liver tumor, occurs mostly within the first three years of life, and while most cases are sporadic and associated with risk factors like low birth weight, preeclampsia, maternal tobacco smoking, TPN, et cetera, one third of these cases do have genetic predisposition, and again, we heard about its correlation with vitamin and FAP and EDWARD and other syndromes. The treatment for most of them nowadays is neoadjuvant therapy in the form of cisplatin and others, and surgical resection with cure rates of almost 70 percent. And we've also heard from a previous speaker that histologically, although the histology classification is quite complicated, but broadly you can think of it as epithelial or mixed. But what's very interesting from our lab perspective is that almost more than 90 percent of these hepatoblastomas have alterations in CTNNB1 gene, which is the gene that encodes for beta-catenin protein. So beta-catenin is a critical component of the Wnt signaling pathway, and just to kind of give a very brief overview, you have either Wnt off or Wnt on. If there is no Wnt present, then beta-catenin is recognized by several of these proteins, including APC, axon, casein kinase, and GSK3 beta. Ultimately, beta-catenin is phosphorylated at specific serine and threonine residues within exon three, and once it's phosphorylated, it becomes recognized for ubiquitin proteasomal degradation. If Wnts are present, then this degradation complex of beta-catenin is no longer active. Beta-catenin is released from its complex in its hypo-phosphorylated form, translocates into the nucleus, and turns on target genes. So in 80 percent of hepatoblastomas, as I had mentioned before, we have activation of beta-catenin, and that is mostly because of two different reasons. One, missense mutations that affect the exon three, so those phosphorylation sites that need to be phosphorylated for its degradation are mutated, or there are deletions in exon three that parts of exon three, which contains all these sites, are missing. And as a result of this, if you do immunohistochemistry for beta-catenin, a lot of the times you'll be able to see very nice nuclear translocation of beta-catenin, and these are two cases. But what was very interesting to us was the work that we did in collaboration with Xin Chen, who's also in the audience in the back, who at the time was at UCSF, but now is at University of Hawaii, we identified that in addition to beta-catenin being present in the nucleus, another protein called YAP1, which is yes-associated protein one, is also simultaneously present in the nucleus of these hepatoblastomas in the tumor cells. So that was quite interesting. The mechanism of YAP activation in these cases is still not known, but it's about more than 80 percent of hepatoblastomas showed this. So to address the causality of these two particular oncogenes in hepatoblastoma, we co-expressed both these oncogenes in the liver using sleeping beauty hydrodynamic tailbane injection, and what we saw was that if you co-expressed these two mutant forms in the liver, within about 8 to 11 weeks, the animals showed significant tumor burden, and they showed, you know, and they died at this stage. These tumors did express the markers of typical markers of hepatoblastoma like MYC, but very interestingly if you do a whole genome sequencing or transcriptomic sequencing, and you compare the overlap of pathways that are activated in our pediatric, in our hepatoblastoma animal model and compare it to the patients with hepatoblastoma, there is a significant overlap in those pathways, suggesting that we have truly replicated a human disease in these animal models. Also if you look at the most up-regulated or down-regulated genes in our hepatoblastoma model in the animals, there is almost 80 percent similarity to the hepatoblastoma publicly available databases of gene expression. So we wanted to further study and try to identify what is downstream of these two oncogenes, YAP and beta-catenin, that could be playing an important role, and this is a work that has just been accepted for publication in American Journal of Pathology and is led by Ed Hurley, who is a neonatologist at Children's Hospital in Pittsburgh. What was identified was that most of these hepatoblastomas expressed this heat shock factor one by immunohistochemistry. When we looked more carefully at the gene expression studies, a lot of HSF1 downstream targets like HSP27, 70, and 90 were significantly up-regulated in these tumors in our model, and very interestingly, if you go and look at it in gene expression studies, in C1 versus C2 form of hepatoblastoma, which is molecular classification, where C2 is a more aggressive form of hepatoblastoma, we found an up-regulation of HSF1 in that subset of hepatoblastomas. And if you look at meta-analysis across about six different studies that had done transcriptomic analysis on various hepatoblastomas, we found, again, significant up-regulation of HSF1 in these patients. So what is this heat shock factor one, or HSF1? As the name suggests, it's a heat shock response transcription factor. Whenever the cells are stressed, HSF1 translocates into the nucleus and binds to its response element on various group of transcription factors and target genes, and then these pathways get activated. You will see activation of specific signaling pathways, which are involved in cytoprotection. They are also playing important role in preventing apoptosis and aging. Over the last almost 10, 12 years, HSF1 has been shown to play an important role in various cancer responses as well. Again, it has been at the center of many different events that are all listed here that are known to be critical players in establishment of these tumors and their progression. So we were pretty interested in trying to pursue what could be HSF1 doing downstream of YAP on beta-catenin in hepatoblastoma. And for that, we decided to look at dominant negative form of HSF1, which is, again, it's a simple plasmid that translocates to the nucleus, but it does not bind to the target gene in the promoters. So the goal was to try and express dominant negative HSF1 while expressing YAP and beta-catenin. And when we do that, what happens is it's a black and white phenotype. These animals are completely prevented from developing hepatoblastoma. The livers are pretty pristine and clear. And only occasionally will you see very, very tiny nodules within the liver. This is also reflected in their liver weight to body weight ratio, which is a good surrogate marker for our hepatoblastoma model. And you can see here dominant negative HSF almost normalizes the liver weight to body weight ratio in these animals. So the question is, what is HSF1 doing in downstream of YAP? And so when we looked at early stages after these dominant HSF1 has been co-delivered and just about two weeks or so, and we start looking at the cells that have taken up beta-catenin and YAP, which can be identified by MYC tag because our plasmids contain that tag. And then you specifically look at V5 tag, which is present on the dominant negative HSF1. You can see wherever this V5 tag was expressed, we saw apoptosis. And when we counted the number of cells, we found most significant increase in number of both T tunnel-positive cells as well as cleaved caspase-positive cells. So what we do think is that HSF1 is very critical in promoting survival of these clones of cells within the liver. And the question, the bigger question is, can HSF1 inhibition impact an already established hepatoblastoma? And that the studies that are ongoing using both genetic and pharmacological approaches, and those studies are now funded internally by the Liver Center Pilot Feasibility Grant, and the initiative is led by Dr. Ed Hurley. But in the last five minutes or so, I wanted to spend time on, you know, can we, instead of targeting what's downstream of beta-catenin and YAP, can we identify ways to inhibit beta-catenin? Because as I said, more than 90 percent of these hepatoblastomas are showing constitutively active beta-catenin, and about 26 percent of hepatocellular cancers in the U.S. and about 38 percent in France have beta-catenin gain-of-function mutations. So we decided to collaborate with different companies, and this is very recent work that is unpublished in collaboration with L. nylum. The goal was, again, to use a hepatoblastoma model here using constitutively active beta-catenin and constitutively active VIAP1, express these two oncogenes, and then use a lipid nanoparticle that they have, which only targets to liver cells and contains an siRNA that is directed against human beta-catenin. And then we gave these, this lipid nanoparticle intravenously every three days for the first about five doses, and then weekly for the next about three or four cycles. And then the animals were euthanized and harvested 24 hours after the last LNP injection for analysis. So what we saw was actually a significant decrease in tumor burden in this hepatoblastoma model as compared to the animals who got either no treatment or got lipid nanoparticle that was just containing a scrambled control. And you can see here by liver weight to body weight ratio that there was a significant decrease in liver weight to body weight ratio in the LNPSI beta-catenin treated group. When we looked at some of the downstream targets here to show, we did find a significant decrease in glutamine synthetase, which is somewhat heterogeneous in this model of hepatoblastoma. It's a very faithful readout in hepatocellular cancers, but in hepatoblastoma, based on whether you're fetal or whether you're embryonal, you can have BGS positive or negative. But we saw almost complete elimination of GS by siRNA to beta-catenin. Myc was a tag that we had used on beta-catenin. And if you look at the levels of myc, you'll find a significant decrease in myc within the tumor nodules that were remaining in these si beta-catenin treated livers as compared to the controls. What was the result on proliferation? We found, again, a notable decrease in cyclin D1, which is an important downstream target of beta-catenin and important for G1 to S phase transition, and hence, one of the factors that drives hepatocyte or tumor cell proliferation. And as a result of this, there was a notable decrease in number of Ki-67 positive tumor cells, which is a marker of S phase of tumor cells. So the key takeaways are beta-catenin activation is frequently evident in hepatoblastomas, and the beta-catenin YAP model represents a relevant model to address biology and test therapies. HSF1 is important in the BY model and in patients, and plays an important role, we think, in regulating hepatoblastoma cell survival, and hence, could be a great target. Lipid nanoparticle carrying si beta-catenin is highly effective in reducing beta-catenin activity, and beta-catenin LNP si beta-catenin decreased hepatoblastoma burden, at least in the beta-catenin YAP model. And I did not have time to present our studies in HCC, but we found an even more profound effect on met-beta-catenin and Nrf2 beta-catenin models, which represent about 20% of all human HCCs. So the exact role of HSF1, I think, in hepatoblastoma still needs to be addressed, and as I said, Dr. Hurley is working on that. Therapeutic HSF1 inhibition is also, testing is also ongoing. And I think we do need some further dose optimization of LNP for si beta-catenin in hepatoblastomas, and how, what exactly, how beta-catenin regulates tumor burden in hepatoblastoma needs to be further addressed through some mechanistic studies. And another major question that we have is, will YAP1 inhibition also be critically relevant especially in hepatoblastomas, just because they are equal partners in that specific type of tumor. So I just wanted to, again, thank some of our collaborators. Xin Chen has been a long-term collaborator. We have a couple of MPI grants together. Ed Hurley is a neonatologist who's really pursuing the HSF story. The study that I showed you with the L-nylem is being done by a very talented MSTP student, Brandon, also helped by Evan Delgado, who's an assistant professor in the lab, and Jun-Yin Tao, who's really the, I think, most talented sleeping beauty transposon individual on the planet. So, and we are really fortunate to have him. And then many others in the lab who are also here and presenting their own work during the course of the meeting. So thank you very much for your attention. Thank you, Dr. Monga. It is my honor to introduce our next speaker, Dr. Jessica Zakman-Rossi, who will be speaking on genomic profiling of pediatric liver cancers. Dr. Jessica Zakman-Rossi. Thank you for the introduction. I will try to skip, first, my disclosure, I will not skip, and speak directly on driver landscape of mutations in cancer liver genes in hepatoblastoma. And I will focus on my talk on the hepatoblastoma. So what we know, as mentioned by Paul and all the speakers, that the Wnt-beta-catenin pathway is really the most frequently altered pathway, both with CTANB1 exome 3. Some of the in-frame deletion were underestimated in the historical series of patients, but in fact, it is very frequent. But some of the non-mutated CTANB1 tumors are, in fact, mutated for APC with frequently germline mutations, and some of them are also mutated for Axin1. The second genetic alteration that is very frequent also in hepatoblastoma in more than 84 percent of the cases is, in fact, the alteration of the chromosome 11p15 that is, in fact, very complicated. But you have to just understand that if you have some alteration in the chromosome 11p15, you duplicate or you have a duplication of the paternal allele that you receive from your father and into the tumor, and this allele coming from the father is, in fact, oncogenic because it induces the expression and the overexpression of IGF2, that is the insulin growth factor type 2, that is really a pro-proliferative factor. So just you have also to remind that the mechanism by which that chromosome 11p15 is altered in hepatoblastoma is a little bit complex, but most of the cases are due to copineutral LOH with a gain of the paternal allele, and then the overexpression of the paternal allele that is oncogenic. What we identified is not only that 84 percent of the cases in the tumor were altered for the chromosome 11p15, but also we found some chromosome 11p15 copineutral LOH, that means really the defect at the chromosome, the short arm of the chromosome 11. Also in 10 percent of the non-tumor liver tissue of the hepatoblastoma, and we know that it is really, this is the example of six, the six first cases that we identified, and you see that we identified this alteration in mosaic, not in all the hepatocytes that are in the non-tumor liver tissue, but only in small proportion of the hepatocytes that are normal hepatocytes, non-proliferative hepatocytes, reaching between 6 percent of the hepatocytes in the non-tumor liver tissue, or more than 50 percent of the hepatocytes in the non-tumor liver tissue. But none of these non-tumor liver tissue were, in fact, mutated for beta-catenin, and when we looked at the tumor, in each cases, we found exactly the same aberration at the chromosome 11p15, meaning that really the tumor is deriving from the mosaic-altered hepatocyte in the non-tumor liver tissue, but then they occur mutation activating beta-catenin, and you see this is all the mutations in each cases that are identified into the tumor. So we can say by that, and we can conclude that mosaic 11p15 alteration in the non-tumor liver tissue is really the early pre-malignant events that occur in 10 percent of the hepatoblastoma patients, and that we validate this result in an extended series of patients recently, and we found that all these patients with mosaic alteration at chromosome 11p15 were, in fact, very young, below the age of three years old, compared to the non-mosaic hepatoblastoma patients without, and children without mosaic at the chromosome 11p15 in the non-tumor liver tissue. The functional consequences of the 11p15 mosaic alteration were deciphered using bulk RNA-seq, and we identified in the mosaic liver overexpression, not only of IJF2, but also other component of the insulin growth factor pathways. We identified an overexpression of progenitor markers, also with hepatocytic stem cells, and also we identified some angiogenic markers. And this is very interestingly shown by our Ph.D. student, Jill Pillay. She showed that the mosaic alteration in the non-tumor liver tissue was, in fact, very complex, and you see, using a dual-color, B-color RNA-scope techniques, that here you have a non-mosaic area in the non-tumor liver tissue of these children, where you have here in blue the mosaic part of the, and so the altered part of the non-tumor liver tissue. So this is really a disease that is a somatic disease in the non-tumor liver tissue, and that is really distributed with a lot of heterogeneity across all the non-tumor liver tissue. Interestingly, what we identified also is that the IJF2 genes that is normally in normal tissue with peripartal area of overexpression, whereas you know very well that glutamine synthetase, this expression is localized at the opposite in zone three of the zonation of the liver around the central vein. In the mosaic tissue, this is completely different, and you see that there is another expression of the IJF2 all over the lobules that, in fact, is disappearing, and also the zonation of glutamine synthetase disappeared. So we diagnosed that using this not only dual RNA-scope, but also with spatial transcriptomics that there is a destruction and alteration of the zonation in the mosaic 11P15 non-tumor liver tissue, and that's lead also with this blue area, and here in this blue area, that are released cluster of and the aggregation of the mosaic hepatocytes that are defined by release overexpression of insulin-like growth factors. So this is interesting because now we can diagnose this mosaic area, and this is a new genetic alteration at the origin of the development of hepatoblastoma. There is also other genes that are accumulated in the tumors this time, and these genes are that are very, there is many recurrent genes that are also altered in the cases of each tumor, so they are very limited for each of them, but some of them are interesting because they could be targetable. For example, RPS6K3 is inducing some arasmide kinase overexpression, IGF19 amplification, and MDM4 amplification could be also targetable in the future. So what about the transcriptomic analysis? In fact, we identified the huge heterogeneity and plasticity of the cells by analyzing more than 100 samples using RNA-seq data. We showed that there is three major subtype of hepatoblastoma, the hepatocytic ones that correspond to C1 in the k-row classification, the liver progenitor ones that correspond to the C2 in the k-row classification, and then the third ones that correspond to the mesenchymal subtype of hepatoblastoma. What is interesting is that not only the samples and different tumors that show this classification, but also within the same tumor, we can have this heterogeneity and the coexistence of different parts of the tumors that could be here hepatocytic or here with the liver progenitor subtype. And in fact, it corresponds exactly to the histological classification with here the hepatocytic molecular subtype corresponding to the fetal very well differentiated subtype of hepatoblastoma, whereas the liver progenitor subtype correspond to the embryonal less differentiated hepatoblastoma. Also, what happened with the immune infiltrate? Immune infiltrate is usually more abundant after the chemotherapy and after the cisplatin exposure to and treatment for the children, but immune infiltrates are never seen really or very infrequently identified in the liver progenitor compartment of the patient, of the tumors. And this is really very significant and showing that really the immune infiltrate is usually more abundant after chemotherapy in the hepatocytic compartment of the tumor, and usually the liver progenitor compartment is a desert of immune cells. But what about mutational signatures that is due to cisplatin treatment? Because one of the hypotheses could be that cisplatin treatment could induce some neoantigenic reaction and then the immune response due to the treatment. But in fact, what we identified is mutational signatures that were characterized by frequent C2T mutation in the specific context and trinucleotide context that, in fact, correspond to really the typical site of DNA adducts that are identified due to the treatment. And this is due to the cisplatin mechanism of anti-therapeutic mechanism of cisplatin. And this mutational signature is, in fact, due to the miss to the reparation and the misrepair of these defects that are due to the adduct to the cisplatin, and inducing a lot of different mutation with this precise signature. The surprise was to identify before, it was not a surprise, because before chemotherapy we never see this mutation, this mutational signature that is termed SBS35, because it's specific of the cisplatin exposure. But after chemotherapy, we identify the SBS45 mutation only in the progenitor subtype of the tumor. And we never identified these signatures significantly in the hepatocytic or mesenchymal subtype of the tumor. And this is also what is very intriguing, because in these cases, for example, with this B-phenotypic tumor, there is no SBS35, no cisplatin signature in the hepatocytic compartment compared to the liver progenitor compartment. So what does that mean? And can we imagine something that is due to the selection after cisplatin? In fact, when we analyze the local metastatic and relapse sample of the tumor in the same children, we identified that in the liver progenitor compartment, there is tons of mutations due to the cisplatin that are subclonal. And then in the metastasis or in the relapse, we have a clonal expansion coming from a single cell with this ton of mutations due to the cisplatin. So it's clearly the metastasis is due to the selection of some cells that are resistant to the chemotherapy. And this is, in another view, how we can represent this result. That means that in a patient, there is some mutations, beta-catenin or MDM4, or alteration of 11p15 in this patient that are common in the primary tumor or in the metastasis. But you see that in the metastasis, there is a huge number of mutations that are due to the SBS, to the cisplatin signature, whereas the cancer-derived gene are exactly the same. The consequence of the, and the reason why there is an accumulation of cisplatin-induced mutation is because of the liver progenitor hepatoblastoma showed another expression of many cancer-derived genes and many genes that are involved in DNA repair. And this is a list of different pathways of the DNA repair that are overexpressed in the liver progenitor compartment. And then, if we analyze also more precisely the different here in this patient, different type of metastasis, M1 to M2 or M3, M prime, and you see that, once again, there is cancer-derived genes that are common, but then you can have different mutations that are induced with a deluge and tons of mutations that are due to the resistance to cisplatin. So that's important because also another student in the lab, Amelie Roring, that identified at the single-cell level that this clonal evolution with longer metastasis was, in fact, due to the proliferation of cells with the accumulation of both specific chromosomal alteration, but also specific mutations that are due to the cisplatin resistance. But in fact, what is important is that in the primary tumor, you have the different component of expression with the hepatocytic component, the liver progenitor component, and here's the mesenchymal one. But in the metastasis, you have also reproducing the three different components, showing that the metastasis is really derived from cell plasticity after a bottleneck of selection of precise and unique cells with the selection of the cisplatin clonal mutations. So this is another representation of different clonal evolution, and you see that it is really reproducible in several patients with a longitudinal follow-up of the patients. The good news is that we identified also some targets and some drugs that could be efficient in a panel of different cell lines, particularly those that are very resistant to cisplatin, and these drugs can counteract the resistance to cisplatin by acting also to the DNA repair. So just to summarize what I showed you, that first we have some new genetic defects that are mosaic alteration at the chromosome 11p15, similar to those that are observed in nephroblastoma also in the predisposition, all occurring in very young children. Then you have the occurrence and the accumulation of beta-catenin-activating mutation or APC-inactivating mutation in almost all the cases. And then you have the cell plasticity with longitudinal heterogeneity, spatial heterogeneity, with specific mutational signatures that are related to cisplatin resistance and accumulated through the misrepair of cisplatin adduct in the liver progenitor subtype of hepatoblastoma. And then cisplatin resistance in all the relapse and metastasis are clonally derived from liver progenitor tumor cells with this cisplatin mutational signature. So this is, they are all also immune-called looking, and this is very important to look at new therapy targeting the DNA repair machinery. So to finish, I want to thank all the people in my laboratory and that participated to this work, and also the funders and the different collaborators. So thank you very much, and now it is time for the questions. Hi, Xinwei, how are you? Very good, thank you. Xinwei Wang from National Cancer Institute. I want to thank the organizer for putting together this very interesting panel. A great lesson for me to learn a lot of new things. I think if it's okay, I can have two questions. One for Dr. Patel, one for Dr. Zuckman-Rossi. So I think the first is a question for Dr. Patel. It struck me quite interesting that you mentioned about the majority of hepatoblastoma patient that you get a biopsy. Now, we know that for adult HCC, biopsy is not recommended in the field because due to the early study or the very few study about the possible tumor seeding with a needle. You know, we know that this actually leads you to really have a much better study of the tumors by taking a biopsy. You know, what are your thoughts on this, whether this can be really implemented into the adult HCC in a second, whether that you have connecting those to the molecular features for your histological feature as well? We have, I cannot cite studies, but we have not seen tumor seeding in the pediatric age group with biopsies. I would say 100% of our hepatoblastomas get biopsied. We typically, they are done by interventional radiology. These are fine needle, they could be fine needle aspirates followed by a biopsy or just a fine needle biopsy, usually 16 gauge. And we request them to just send us multiple course. They do multiple passes many times. And in fact, we actually had another separate study documenting how on some of the lesions passes that were lesional, lesion to transition, and normal were also important. And I mean, I don't have data, but none of those patients had any issues with seeding of tumors. Yeah, so just to clarify, I think some of the plan for the adult HCC for the tumor seeding is very based on very limited data. And that was the only justification for the recommendation in the field to move forward. So I think it's time probably to discuss a little bit more about the biopsy for adult. So let me ask a question for adult. But maybe I can complete also that in many cases, there is some depth into the diagnosis as you mentioned also with the histology. And when you are dealing also with young children, you cannot have a depth, but in fact for adult also. So, but I think that it is remandatory to have a firm diagnosis and not only by imaging. And also the subtype in America is, so when you have a very well differentiated tumors, in America you have surgery first and without chemotherapy. And usually in Europe, you have neoadjuvant chemotherapy before the fit protocol for all the patients, and then surgery, and then adjuvant chemotherapy. So it's completely different protocols compared to adult. But we have also biopsies and in fact samples that are coming from the surgery. And most of the samples are very well analyzed after surgery. I agree. And it actually gives a very unique opportunity to study pre-therapy, post-therapy samples as Dr. Rossi showed. So question for Jessica. So you demonstrate that obviously the hepatoblastoma is also quite heterogeneous if you look at the single cell level, particularly plasticity is evident. But it's similar to the adult HCC. Yet adult HCC tend to have a worse outcome than hepatoblastoma. Do you have any thought whether they are different in terms of heterogeneity? Yeah. In fact, in children, it means the hepatoblastoma are really much more first heterogeneous and then plastic, demonstrating cell plasticity. We were very impressed. I mean, usually we were working with adult hepatocellular carcinoma, and in hepatoblastoma it's completely different. But what was really very striking for us is that the correlation also with the histology and the compartmentation of the tumor is very, very important. So the question now is to really understand why we have this type of compartmentation into the tumors. And I think you showed also some pictures where you have really the area that is hepatocytic or fetal hepatoblastoma completely separated from the embryonal one with really a clear boundary between the two, and to understand how the transition occur, whereas it is the same cancer of the gene origin, is really under question at this moment and very important to understand. I really agree with you on that. And it takes true pathology, molecular biology collaboration, because once you have a sample that you're analyzing for genetic alterations, it's important to know where the sample has come from, you know, which histologic area is represented by that sample. And that can definitely be done through micro and macro dissections, but a blind study of samples will not be useful moving forward. Thank you. This is Samar Ibrahim from Mayo Clinic. Thank you for the outstanding talk. I have a quick question for Dr. Manga, and it might be naive. So I was wondering, like, inhibiting beta-catenin improved the hepatoblastoma tumor burden and all the markers, and I was wondering if you can target this upstream with a destruction complex. We know JSK3 inhibition is beneficial in pancreatic cancer. Is this, you know, can this be? So, theoretically, I don't think it will work, but practically, you know, we haven't tried it. I think it will be nice to try a porcupine inhibitor or some inhibitor that is upstream and affects Wnt release from a cell. We think the mutations are gain of function, they are downstream, so anything upstream of it is not going to work. So once beta-catenin is in the nucleus, then I think you need something to either inhibit its complex with TCF or CBP, or you have to just shut down the gene expression of the gene itself. Okay. Thank you. Yeah. Thank you. Hi, Shan. I just have a quick question for the panel. We know that immunotherapy has been taking the oncology, you know, field, and, you know, hepatoblastoma, you know, like, has beta-catenin mutation, which has been proposed to be maybe linked to a, you know, resistant to immune checkpoint inhibitors. Is there any evidence that this is really the case? Has immunotherapy been applied to, like, hepatoblastoma? What is current status? Anyone have any thoughts? So to the best of my knowledge, no. I think children are a very unique set of population, and I think immune checkpoint inhibitors have to be really careful. And I'm not really sure if the neoantigen load, et cetera, in a pediatric tumor, which is very low mutational rate, you know, whereas HCC, you can see up to 100 gene, you know, mutation that can eventually occur, which will alter the overall differentiation status and the, you know, the neoantigen load. I don't think that happens in hepatoblastoma. But again, I think these are questions that would need to be addressed directly. And we haven't given a PD-1 inhibitor or PD-L1 inhibitor to our YAD-beta-catenin model, but can be done. Yeah. But maybe I can complete also that in the liver progenitor subtype of hepatoblastoma, you have a ton of mutations. Some of them are generating or supposed to generate some neoantigens, because using the whole genome sequencing, you have more than 10,000 mutations due to the cisplatin signature. And this is a dessert, in fact, naturally, after there is no immune infiltrate. So it seems that it's more complicated than in other type of tumors, because also of tumors in very young children, and we need to have more data on that. Can I ask one question? Yeah. For Dr. Rossi. So this is about the SBS-35 mutations that you showed in the beautiful data. Have these mutations been reported in other pediatric embryonal tumors or any other tumors? Because we do see this pattern in other tumors, where a subclonal evolution happens, and then even at the site of metastasis, there is plasticity, right, like neuroblastoma. Yes, it has been reported in other type of tumors that has been treated by cisplatin, for sure. And also in adult, in some adult cases with hepatocere carcinoma treated by TACE with cisplatin, that is quite not usual, but can happen particularly in Japan, for example. When you look at ICGC data, for example, you see some SBS-35 mutations, mutation signatures also in these tumors of patients treated by, yes, that are supposed to be treated by cisplatin. But, yes, it's a common signature that is due to the cisplatin and the resistance, and in different type of tumors. Thank you. So, I know this session is focused on liver cancer, but I'd like to pose a question to Dr. Wangstein. You know, given the very early diagnosis of hepatoblastoma in children, particularly the younger 11p15 mosaics, you raise the issue that there was a 5 to 15 percent risk of familial adenomatous polyposis, which is not an insignificant risk. Should we, and is there any clinical guidance or recommendations regarding how soon we need to be screening these patients who have been diagnosed with hepatoblastoma, say, as early as 18 months or even younger? Well, thank you for that question. Again, I'm an adult gastroenterologist, but in my reading, I don't think this is reaching, you know, the level of, like, consensus statement or things like that, but expert opinion is to perform germline genetic testing for APC. I just wonder, from the pediatric hepatologists in the crowd, who routinely screens colonoscopically for FAP in your newly diagnosed HB patients, either pre-transplant or post-transplant? Or does any molecular testing? Right, so it's an interesting association, and I'm just curious if our pathologists who may see these cases down the road can maybe comment on some of these occult findings that the clinicians may not be aware of. Hi, how are you? I'm Allie O'Neill. I'm a pediatric oncologist, so not a hepatologist. But the one thing I can add is at least we're fortunate to routinely screen these kids, and when they have either a variant of unknown significance or a true APC gene mutation, we do recommend they get colonoscopies as soon as 10 years of age. So I think from the pediatric perspective, we are, you know, funneling them towards our GI colleagues at an earlier timeline than we would otherwise have, and that's independent of family history, not really knowing how to think about an APC germline in these patients that already have a diagnosed malignancy at a young age, if that helps. It helps a lot because, you know, I'm sure most centers have dedicated FAP clinics, but knowing that we can wait a decade before screening these patients for FAP, it makes me a little nervous knowing that there's a 5% to 15% risk, you know? So in France, maybe the situation is a little bit different because there were some suggestions, strong suggestions, to test for beta-catenin into the tumor, and then in the tumors that are not mutated for beta-catenin to propose a genetic testing for APC. But, in fact, it was three or four years ago when we started this program. Now the things are different because also the cost of whole exome sequencing and whole genome sequencing is decreasing a lot. And the idea is because this disease is very rare, and you have also other genes that are mutated, and also for epithelial carcinoma, children, and et cetera. So overall, with liver disease in children is to propose maybe in the future, and we try to push for that, some genetic counseling and sequencing using whole exome sequencing and whole genome sequencing in the blood, but also in the tumor because we need also to make some progress into research and to identify some new therapies and targeted therapies. I remember one case in our practice where we saw nuclear staining for beta-catenin in the non-tumor liver, which was unusual and difficult to explain, led to further investigations, led to a germline identification of a germline mutation in APC in that case. So I think if you see that finding in the non-tumor liver, that can maybe take you in that direction. Richard Thompson from King's. Just to say that now in the U.K., it's routine for all pediatric tumors, including hepatoblastomas, but all pediatric tumors get germline and tumor whole genome sequencing as part of clinical care. So obviously we report somatic mutations, but we also report germline mutations in every cancer. Thank you. Richard? I think along the same lines, maybe even looking for SNPs and PNPLA3, which are known disease modifiers in liver disease, would be relevant because especially of the polygenic. And I'm looking forward to David Kaplan's study that is going to be presented. So hopefully there are messages to be observed in that situation as well. So for PNPLA3 in adults, I fully agree. I think that also as you mentioned, I think it's very interesting to look at mutations with low penetrance and with low functional consequences. And that's true also for APC and maybe for other genes. We don't know exactly which one before having all the screening and the identification of rare variants that are associated with predispositions. So I don't know if you have any comments on that. But there is also the story of alpha-1 antitrypsin mutations with low penetrance of the phenotype. We have also some glycogen storage disease with low penetrance of some alleles and mutations. And different genes now that accumulate mutation and with low penetrance variants that are probably underestimated and probably could also contribute to familial cases. Yes, I completely agree. I mean, I think this is a super interesting conversation. I think the conclusion should be that these cancers are very rare in the pediatric population. And if molecular testing is available, then you should take advantage of that. But it's still not clear how many genes to test and what the significance is. For things like PNPLA3, the risk allele is very common. I mean, basically, like 40% of us have in this room probably have the risk allele. So for an individual with, like, a cancer, that makes very little difference. But it's possible, I think, that in the future we will have a polygenic risk score for patients that have cirrhosis that are at risk for HCC. And we may be able to kind of funnel them to a different type of screening for HCC based on that kind of score. Dr. Monga, your mouse work and the summary of the significantly reduced tumor burden in inhibiting heat shock factor one and beta cat and YAP seems to hold a lot of promise. Any chance these therapies may reduce the need for our pretext for hepatoblastoma patients to avoid liver transplant? That is the goal. That is the goal. And I think we were, you know, pretty excited about four years ago when we almost thought that we'd have a drug. But then, you know, when you depend on collaborations with companies, the priorities change and things change. And that was exactly what happened. But we have a renewed interest from a specific company that I shared. They have their reagent works really well. And the delivery is meant for liver cells and liver tumor cells. So I think there would be specificity. We always worry about global beta catenin inhibition or YAP inhibition in a pediatric scenario. But this would be a very liver centric. So we do think and we are hoping to have some safety studies starting as early as next year. Thank you very much. If there is no other question, this is time for the closing remark. I think we have learned a lot on the mechanism of hepatoblastoma particularly, but also in hepatocellular carcinoma and with the predominant activation of beta-catenin in all these tumors. Remember also with the adenoma. And so the spectrum of tumors and the phenotype of the tumor is really very broad and different according to the different tumors. But sometimes with the same mutation at the origin that is beta-catenin that is really central amongst the oncogenesis in children. But then we will move in one hour in the second part of the session. So at 3.30. And we will have a focus on the fibrolamellar carcinoma because we didn't discuss a lot. We will have a presentation by Simon Sanford and we will learn much more on this very specific subtype of hepatoblastoma with promising candidate therapies. And we will learn more with Dr. O'Neill with the different systemic therapies that we can apply to the children and where is the hope to cure liver tumors in general in children. That's also very important to don't forget surgery. And there is also some innovation in terms of surgery. And we will learn more with Gregory Thiault with novel approaches. And then we will learn also from the patient side what are the comments on the progress that remain to be done in terms of research but also in terms of discovery of treatment and to improve the efficacy of the treatment in liver cancer patients. Thank you very much and see you in one hour.

pediatric liver cancers

hepatoblastoma

histological subtyping

genetic testing

diagnosis

targeted treatment

biopsies

Wnt-beta-catenin pathway

11p15 alteration

tumor microenvironment

HSF1 inhibition

immunotherapy

germline genetic testing