From myself and Joseph Lovett, Chair of the Epidemiology and Neoplasia Special Interest Group, it is my pleasure to introduce the moderators and speakers for the second part of our virtual session, Hot Topics in the Pathogenesis and Treatment of Liver Cancer. The second session will include four 20-minute lectures, followed by a wrap-up by Professors Jessica Zuckman-Rossi and James Trotter. Our session moderators will be Professors Lovett and James Trotter. The first talk, Trial Design and Endpoints in Advanced Hepatocellular Carcinoma, Overall Survival and Progression-Free Survival, will be given by Joseph Lovett. Our second talk, Immunotherapy for Hepatocellular Carcinoma, Current Results and Need of Combinations, will be given by Richard Finn. Our third talk, Advances in Precision Oncology in Intrapartic Cholangiocarcinoma, will be given by Greg Gores. And our last talk of the session, Role of Molecular Tumor Boards for Management of Liver Cancer, will be presented by Robin Kelly. Without further ado, I turn you over to our moderators for the session, Joseph Lovett and James Trotter. Good afternoon. Welcome to the Joint Hepatobiliary Neuroplasty, HIG and Liver Transplantation and Surgery, HIG Symposium. We're giving you the welcome and introductory remarks. I am Dr. Jean-Marie Lovett from Mount Sinai, and I'm co-chairing these introductory remarks with Dr. Michael Shilsky. The topic of the session is on hot topics in the pathogenesis and treatment of liver cancer, both hepatocellular carcinoma and intrapartic carcinoma. Here you have the outline of the session. We'll have two parts. Part one will be addressing the concept of pathogenesis and treatment for early tumors. And we will have two moderators, Professor Jessica Zucman-Rossi and Professor Michael Shilsky, and three presentations. First one, pathogenesis of liver cancer, targeting the FGF MAP kinase axis in the basal carcinoma and intrapartic cholangiocarcinoma, will be addressed by Sandra Rebusio. Second presentation on resection versus transplant for the management of intrapartic cholangiocarcinoma by Julie Heimbach, and the third presentation addressing the question on are there any improvements in the management of HCC with local-regional therapies by Professor Laura Cooley. The second session will be devoted to trial design and management of advanced tumors. The moderators will be myself and Professor James Trotter. The first presentation will be focused on trial design and endpoints in advanced HCC, overall survival and progression-free survival, and I will address this topic. Second presentation will be on immunotherapy for HCC, current results and need of combinations by Professor Richard Finn. Third presentation on advances in precision oncology in intrapartic cholangiocarcinoma by Professor Greg Gorse. And finally, last presentation on the role of molecular tumor boards for the management of liver cancer by Professor Robin Kelly. Finally, we'll have a wrap-up and closing remarks that will be addressed by the co-chairs, Professor Jessica Zucman-Rossi and Professor James Trotter. I'd like to thank all of you for attending this session, and I am hoping that you will enjoy all these hot topics. Thank you very much. Good afternoon. I'm Dr. Jose Maria Llobet, Professor of Medicine at Mount Sinai, New York, at the University of Barcelona. I'm going to talk about trial design and endpoints in advanced HCC, OS, versus progression-free survival. Here you have my disclosures. And here you have the outline of my presentation. We'll talk about basic concepts in trial design, then standard of care in advanced HCC, trial design in phase 2 and in phase 3, endpoints with the debate OS and progression-free, and also objective response rate and resist and modify, resist, and finally, and modify. Recently, the ASLD panel of experts on trial design in HCC, we have published the paper trial design and endpoints in hepatocellular carcinomas, a consequence of the ASLD consensus conference. So, here you have the members of the panel, chaired by Dr. Luis Rovers and myself. I want to thank all of them for their participation in this important panel. We presented a set of recommendations in terms of target population. Generally, we have to select according to BCLC, child PUC classification, mostly child PUC A, also consider aldegrade and melt the skull in some circumstances, and in terms of selection of endpoints, I will provide some granularity afterwards, but the primary endpoint for phase 2 or phase 3 studies is OS. Progression-free is primary endpoint for phase 2 studies, assessing primary treatments in intermediate and advanced, and can be considered also in phase 3 studies, intermediate and advanced, with strict rules for calling superiority. Recurrence and survival is the primary endpoint in phase 2, 3 studies assessing adjuvant treatments, and then we have objective response that is recommended to be assessed in phase 2 and in phase 3 as a secondary endpoint, and patient-reported outcome was also recommended as a secondary endpoint in phase 3, and composite endpoints are coming, OS plus progression-free, and CDER. Regarding the control arm, in first line, we recommended Atezolizumab plus Bevacizumab, and in second line, Regorafenib, Cabon, Ramon, I will explain you about Sorafenib and Lembatinib as well, and also here you have the list of stratification factors in front line, ECOF performance status, macrovascular invasion, extravagic spread, FP more than 400, geographic aversion, and etiology in case you are testing Sorafenib, and in second line, mostly, we divided macrovascular invasion and extravagic spread as independent prognostic factors. Also the type of progression. What is the standard of care for advanced HCC, and I'm posing this issue because it's critical for trials, and we know from the trial, SHARP trial, standard of care was Sorafenib, this was very clear. All the trials afterwards in front line have been compared with Sorafenib, and here you have the list of trials, front line, and in second line. The first drug or combination that have shown superiority to Sorafenib has been Atezolizumab plus Bebacizumab in this phase 3 trial reported in New England this year. The combination was superior to Sorafenib in terms of overall survival with a hazard ratio of 0.58, and a p-value of 0.0006, and this lead to stopping the trial as a result of superiority. Another drug, Limbatinib, was designed in a phase 3 trial with a non-inferiority design compared to Sorafenib, and here you have hazard ratio 0.92, there were no difference between Limbatinib and Sorafenib, but hit the non-inferiority end point, so it's similar to Sorafenib. A different case is Rigorafenib, that was the first drug approved in second line. In this case, the trial design was comparing Rigorafenib versus placebo in patients that tolerate Sorafenib in front line and progress to this drug. As you can see, the hazard ratio was 0.63, Rigorafenib was superior to placebo in second line. In this case, we have Cabozantinib versus placebo in second line advanced HCC, again in patients progressing to Sorafenib, this was also positive with hazard ratio 0.76. And finally, we have Ramucirumab, that the phase 3 RIG trial was negative with a subgroup analysis positive for FP more than 400, and then the RIG tool was actually positive versus placebo, and I'm showing here the results of the meta-analysis, where median survival 8.1 months for Ramu, 5 months for placebo, hazard ratio 0.69, so all these drugs were approved in patients progressing. So when we analyze all the trials in front line, as you can see here on the left-hand side, we can see that Ateso plus Vepacizumab has hazard ratio 0.58 with a 95% confidence interval, the upper boundary is 0.79, far below the one that defines superiority. So it's the only drug superior to Sorafenib. And then we have Lemvatinib with hazard ratio 0.92, upper boundary 1.06, that despite that crosses the one, is within the area of non-inferiority, since the trial was designed for non-inferiority, the trial was positive, and Lemvatinib is non-inferior to Sorafenib. Conversely, Nivolumab, the trial was designed for superiority, and therefore the trial was negative because it was crossing the one. In the other side, in second line, we have three drugs with a positive trial, Regorafenib, Ramucirumab, and Cabozantinib, compare all of them to placebo, and all of them in second line following progression to Sorafenib. So this will be the summary of the current management of HCC and the rationale to choose a control arm in a randomized study. Front line is a tezolizumab plus bevacizumab, however, certainly a percentage of patients will not be able to receive this combination, either because they have any contraindication for any of the compounds, or they have autoimmune diseases, or they present viruses untreated, or either other indications or other contraindications. So result of that, Sorafenib or Lemvatinib can still certainly be used in front line, or nonetheless can be used in second line following progression to a tezolizumab. And then we have the drugs Regorafenib, Cabozantinib, and Ramucirumab that has shown superiority to placebo progressing to Sorafenib will be used after the progression to these drugs. Let's move to phase two proof of concept studies and accelerated approval. These schemes shows you the classical phase one, phase two, and phase three mode to get to regular approval. However, in oncology, accelerated approval is very common based on phase one B, phase two data with very potent signals of efficacy, and certainly more than 30 drugs have been approved by FDA based on this concept. This is an article that summarized accelerated approval in oncology between 1992 and 2010. And as you can see, 47 drugs were granted accelerated approval based on an endpoint of response rate 33, time to event endpoint nine, and six other endpoints. However, not all of them were finally confirmed at the phase three level. As you can see, 55 of these drugs and trials got regular approval based on phase three. Unfortunately, 45 were withdrawn either because of negative results or incomplete randomization. What is the implication in trial design in HCC of this feature? Well, first of all, the panel recommends large single arm phase two trials with a strong signal of efficacy may suffice to justify phase three. Of course, it would be ideal to have randomized phase two studies in all the cases. And the threshold of the signal is something that is not clearly established, but generally in terms of objective response should be certainly beyond 20% of objective response. On the right hand side, I'm showing five examples of drugs tested in phase two that either achieve breakthrough designation or were approved by FDA, Nivolumab, Pembrolizumab, and Nivolumab plus Ibilinumab. The three drugs or combinations are available now in the US. Then Lemvatinib Pembrolizumab got breakthrough designation, but was not approved and pending the phase three and Fizogatinib did not get any of those. These are the results of the check main of 40 and as you can see, a small percentage of patients achieve complete or partial response with an outstanding outcome. And also patients with a stable disease show median survival of 16 months. The objective response was not associated in this study with the PD-L1 status at baseline. And these results lead to approval for Nivolumab in the second line and provide the rationale for the checkmate 459. This study, 15% of the patients achieve objective response following Nivolumab. However, in terms of the primary endpoint of overall survival, Nivolumab did not show superiority compared to Sorafenib. Similar situation happened with Pembrolizumab versus placebo that was approved in second line in HCC by FDA and justify the keynote 240, randomizing the patients to receive Pembrolizumab versus placebo in second line. Pembrolizumab lead to objective response in 18% of the cases. However, the overall primary endpoint, overall survival with a pre-specified P-value of 0.0174 was not achieved, was not hit and therefore, despite the clear trend, Pembrolizumab was not declared superior to placebo in second line. Finally, the combination of Nivolumab, ipilimumab has been tested in the setting of phase two investigations and the Nivol-EP with 50 patients, the first combination achieve objective response in 32% of the cases with a median duration of response of 17 month and median survival of 22.8 month and this was seen by FDA as a signal of efficacy that granted accelerated approval in second line. Another combination promising Limbatinib plus Pembrolizumab, this is a phase 1b study including more than 100 patients that with this combination achieve around 40% objective responses by modified resist and up to 86 disease control rates. In terms of progression-free survival, for this cohort was a median of 9.3 month and overall survival 22 month. As a result of that, a breakthrough designation was granted by FDA for this combination. Finally, what about specific precision medicine in HCC? Well, the most common mutations in red are known that are undruggable, whereas the uncommon mutations generally below 5% prevalence are some of them druggable. It has been established that 25% of HCC tumors have at least one actionable target and this small percentage compared with the performance of other solid tumors that have up to 60 to 80% of actionable targets, for instance thyroid or melanoma tumors. This study conducted at Memorial Sloan Kettering Cancer Center with 10,000 patients, 500 mutations were checked and as a result of this, 37% of the patients received a specific TKI or monoclonal antibody directed to these mutations. When we checked the performance of liver cancer, 45% of biliary cancers received precision oncology as opposed to only 5% of HCCs. This is the case of FGF19, a known oncogene in HCC, amplified in 7% of the cases and upregulated in 20% of the cases. In the preclinical models, PDX models, the drug blocking FGF receptor 2, here shown in blue, is both effective in FGF or overexpressed and FGF amplified tumors in PDX. We also test this drug, Fisogatinib, in phase 2 study in patients with FGF19 immunostaining positive and immunostaining negative, and certainly the objective response rate in patients with FGF19 positive was 16% significantly higher than patients with FGF19 negative that was 0% proving the concept of that Fisogatinib is able to block the signaling cascade. Let's move now to criteria for phase 3 design and regular approval. I already mentioned the selection of the target population and the control arm. And now let's focus on the expected outcomes. As you can see here for frontline, a tesolizumab plus bevacizumab we're expecting median survivals beyond 17 months with sorafenib 14 months and lembatinib 13 months. And in patients progressing to that with regorafenib 11 months, cabozantinib 10 months and ramucirumab in patients with FP more than 400 we're expecting a median survival of around eight months. This table summarize you the variables that the panel thought that should be included in any clinical trial in HCC and disclose the characteristics of demographics, tumor description, staging system, liver function and general health, including ECOG performance status. In terms of stratification factors prior randomization in first line and in second line are quite similar. In first line ECOG performance status macrovascular invasion, extrabiotic spread together FP more than 400, geographical region and eventually etiology if we're testing sorafenib. Whereas in second line, macrovascular invasion, extrabiotic spread should go separate and also type of progression might be considered. Let's talk about endpoints overall survival and progression-free survival. Overall survival is the most important endpoint in phase two and particularly phase three studies assessing treatments in intermediate and advanced HCC and progression-free survival is the primary endpoint in phase two assessing primary treatments intermediate and advanced and can be now considered co-primary in phase three studies in intermediate and advanced with the strict rules of calling superiority. The magnitude of benefit for overall survival has not been established but threshold of hazard ratio below 0.8 is sound for capturing the benefit for patients in advanced HCC trials. We conducted on a study to understand the role of progression-free survival and the association with overall survival and we use 21 randomized control trials conducted in advanced HCC, 12 in front line and nine in second line that are listed here. There was a clear significant correlation with an R of 0.84 and R square of 0.71 between progression-free and overall survival and interestingly enough, as you can see here in the box, those are trials showing a hazard ratio for progression-free survival below 0.6 show all of them and they are in green overall survival differences. After the publication of these, five additional trials have been reported and two of them were positive for the primary endpoint of overall survival, the InBRAVE-150 and the SORA-HIEC and both of them show a hazard ratio for progression-free survival below or equal 0.6. Conversely, three additional trials were negative including the PRODIGE-11, the CHECKMATE-459, the KEYNOTE-240 and all of them show a progression-free survival hazard ratio above 0.6. Whether this threshold that has been established in advanced HCC based on now 27 randomized control trials might be applied in intermediate HCC, including CASE or even in the new trials or advanced HCC, we will see which are the results but certainly most of these trials, progression-free survival is currently either the primary or co-primary endpoint in most of them. Finally, I want to talk about objective response and time to event resist versus modifier resist. We recently reviewed the M-resist criteria for HCC and certainly improve some novel refinements and check for the performance. As you know, the main difference relies on the description of response that in M-resist is defined by the diameter of the uptake lesion at the arterial phase, whereas in modifier resist is the baseline and after treatment diameter, plane diameter, regardless of the uptake. In terms of definitions of complete response, partial response, stable disease and progression, they are identical for both. This table, I'm summarizing the objective response rate described in intermediate HCC patients treated with CASE and objective response ranges between 41 and 62%. And in this meta-analysis of seven studies testing CASE, including 1300 patients, there was a significant correlation between objective response and overall survival with hazard ratio of 0.38 compared with patients not achieving response. In frontline HCC, there is always a difference between objective response by modified resist compared to resist, in this case, 24% versus 18%, 9% versus 6% and so on and so forth. And in four studies has been shown that objective response predicts overall survival with modified resist. Finally, summary of unmet needs captured by the panel. This table summarizes the unmet needs, adjuvant therapies after resection local ablation, neoadjuvant prior resection liver transplant, local regional or SBRT enlarging on surgical tumors, combination with chemoembolization in intermediate, combination or triple therapies in frontline, immune-based therapies for patients with HCC and imperliver function, CHALPUG-B, pivotal proof of concept phase two trials, enrichment with oncogenic drivers, systemic inclusion of cost benefit analysis, identification and validation of biomarkers, inclusion of quality of life, patient reported outcomes in the study, patient reported outcomes in the trials, and also integrate molecular subclasses into the clinical staging system in order to better guide treatment allocation. So in conclusions, the panel of experts of SLD in trial design and endpoints define the target population in frontline as BCLC-B patients progressing to local regional therapies or BCLC-C patients, CHALPUG-A ECOF performance as zero to one. Control arm in frontline should be a tesolizumab plus bevacizumab and expected OS is 17 mark. And in other lines can be sorafenib, lembatin first, and then rego, cabo, and ramo. Regarding type of studies and endpoints, the double blind trials are recommended and better than open-label trials. Accelerate approval is feasible in HCC, generally the signal of efficacy in terms of objective response is around 30% and median OS around 20 mark. And in phase three, the unquestionable primary endpoint is OS, but progression-free survival has shown a clear correlation with OS and is recommended as a co-primary, particularly if the hazard ratio for progression-free survival achieved is below 0.6. Measure of response, certainly modify resist and resist are useful tools. And among the main needs, patient reported outcome should be included in the trials. And we recommend mandatory biopsy and liquid biopsy for clinical trials and clinical studies. Thank you very much for your attention. Hi there, I'm Dr. Richard Finn from the Geffen School of Medicine at UCLA. And I really appreciate this opportunity to join you at the AASLD virtual meeting to talk about immunotherapy in liver cancer. These are disclosures. So when we're talking about immunotherapy, as it stands today, the most advanced drugs in development are immune checkpoint inhibitors. Clearly, there's other strategies being looked at, but for the sake of time today, we'll focus on the data that's been generated with this class of drugs, which includes anti-CTLA-4 antibodies, as well as monoclonal antibodies against PD-1 or PD-L1. Keep in mind, these targets, CTLA-4 and PD-1, really are designed to down-regulate the immune system. And in the context of malignancy, this is harnessed to allow cancer cells to escape. CTLA-4 playing an important role in the priming phase of T-cell development, and then PD-1 and PD-L1 being important in the effector phase. Now, the first drug of this class to be looked at in liver cancer was tremulumab, a CTLA-4 antibody. This study, led by Bruno Sangro many years ago now, looked at single-agent tremulumab in a population of advanced liver cancer patients who had previous treatment, both child PUE A and B, looking mostly at safety and then also some efficacy. Now, first to define the tolerability, while it did generally appear well-tolerated, there was this increase in liver enzymes, which put somewhat of a hamper on the further development of tremulumab based on this study. And again, it was a relatively small cohort of patients. With that being said, there were some responses. The overall response rate was 18%, disease control rate of 76%, and a TTP of six and a half months and eight months, which now, knowing what we know about advanced liver cancer in the second line, might've been a signal. And also interesting enough, and before the large use of the DAAs, there was a decrease in the viral load. The first drug to be approved of this class in advanced liver cancer was nivolumab. And this was based on the Checkmate 040 study, which was a single arm study geared at looking at the safety of nivolumab in an advanced liver cancer population of patients, again, child PUE A, and initially in a dose escalation phase, looking at various etiologies of liver cancer. On this waterfall plot, you can see that regardless of etiology, there's a subset of patients who derive significant benefit with responses. And this was the basis for the accelerated approval of nivolumab with a blinded independent central review response rate of 15% in 154 patients in an escalation and expansion phase in the second line. And duration of response was over 16 months. Now, the challenge here is we don't have a biomarker to know who these patients are. If we look at PD-L1 expression, it does not capture perfectly the patients who respond because there's patients who do not have PD-L1 expression who still respond to the drug. Now, the drug did not have any new safety signals. Actually, the treatment-related adverse events were generally manageable. You can see here relatively low grades of liver enzyme abnormalities. And really there was nothing seen in the liver cancer cohort that was not seen in prior cohorts of other malignancies. Now, the confirmatory study was CHECKMATE-459. This was an open-label study of single-agent nivolumab versus serafinib, a phase three study in patients who are child-PUA and treatment-naive. Now, this study read out and was presented at the ESMO meeting in 2019. And as you can see here, overall, this was a negative study. Serafinib had the longest survival in any of its phase three studies with survival of just under 15 months. Nivolumab actually had a quite long survival of over 16 months. However, the hazard ratio was 0.85, the confidence interval was over one, and the p-value was not significant. The same can be said for the progression-free survival in that this was not a significant difference. But it's been thought that perhaps checkpoint inhibitors really do not separate the median as much as develop a tail to the curve. And median PFS might not capture that. Now, this study did confirm the activity of nivolumab in advanced liver cancer. The overall response rate was 15%, very similar to what we saw in the single-arm phase two study. However, the decontrol rate was not significantly different from serafinib, 55% versus 58%, largely driven by a number of patients receiving stable disease with serafinib, which is how we understand the drug works. The other thing that could complicate the frontline endpoint was subsequent therapy. Overall survival captures not only survival from the frontline treatment, but also subsequent treatments. And you can see that many patients on serafinib arm went on to receive immuno-oncology agents, 20% of them. And also a number of these investigational agents were also immunotherapy-based. And this became access because these drugs were approved in several countries in second line, the IO agents, but also in the context of clinical trials. Nivolumab was very well tolerated, right? If you look here at the side effect profile, the lighter color bars are low-grade adverse events, the darker are higher grade events. And you can see diarrhea and hand-foot syndrome, things that make giving serafinib challenging occur with very low incidence with nivolumab. And other side effects that occur more common generally can be managed with supportive care. Now, the same data set exists very much for single-agent pembrolizumab. That is to say in a phase two study of various etiologies, response rates were around 15 to 20%, regardless of etiology. That led to this study, which was a confirmatory phase three study in second line. It was pembrolizumab versus placebo in child PUA patients with two co-primary endpoints of overall survival and PFS. This study confirmed the response rate of pembrolizumab, 18% with a long duration of response over a year, including a subset of patients who have CRs. And if we look at one primary endpoint, overall survival, you can see that the curves separate. 10 and a half months in the placebo arm is the survival, median survival of 14 months with pembro. Hazard ratio was 0.78. Upper limit of the confidence interval is less than one, and the p-value is 0.0238. However, we cannot declare this study a positive study because the pre-specified statistical plan, which took into account two co-primary endpoints of PFS and OS, as well as several interim analyses, required that the alpha be 0.0174 to be positive. Needless to say, this study, like Checkmate 459, confirmed the activity of single agent ED-1 inhibitors. If you look at time to progression, median TTP was significantly improved, and you can see that there is a tail to this curve at 20 months, I'm sorry, at two years, about 20% of patients still have not progressed with emerylizumab. Now, why did the single agent studies fail? Well, one could argue that it was the impact of second line treatments or treatments beyond progression. The other thing is that the response rate of 15 to 20% was not high enough to beat the effect of serafinib. And that to improve this, we need either a biomarker to identify the patients who benefit or look at combinations that improve the response rate. So one of those approaches is to combine PD-1 inhibitor or PD-L1 inhibition with anti-VEGF therapy. Bevacizumab is a monoclonal antibody to VEGF, which has been approved for many years in oncology in combination with chemotherapy. And initially the feeling was that drugs like Bevacizumab normalize malignant vasculature and therefore starve a tumor of blood supply. However, our understanding now has matured to understand that as the vasculature normalizes, it does allow pro-inflammatory cells into the tumor microenvironment. And if we could combine that approach with a drug like atezolizumab, we could reverse the negative feedback on the immune system and allow tumor, anti-tumor activity. Now, the IAMBRAVE-150 study was published in the New England Journal of Medicine in May of this year, and the regimen was actually approved in June. This is the study design, an open-label study of atezo and Bev given IV every three weeks versus serafinib at its standard dose with two co-primary endpoints of OS and regression-free survival by independent review. Again, concentrating on the child PUA population, and importantly, patients in this study were required to have an upper endoscopy prior to coming on study, at least within six months, given the risk of bleeding with drugs like Bevacizumab that affect the VEGF axis, and taking into consideration a cirrhotic population that might be at risk for variceal bleeding. Now, here are the baseline characteristics, very typical for a liver cancer population, majority are male, majority of patients in this study came from outside of Asia by design, child PUA, and about 80% of patient had Barcelona C advanced, but actually about 15% of each arm were intermediate patients who were not candidates for TACE or progressed after TACE. Here, you see other baseline characteristics which were generally well-balanced. You can see here that over 70% of patients in each arm had macrovascular invasion or extrapatic spread, and all the etiologies of liver disease were represented. This is the primary endpoint, the study was stopped at the first interim analysis given the benefit seen, and therefore the median follow-up at this time was only about eight and a half months. But needless to say, Serafinab had a survival of 13 months, Nitesilizumab and Bevacizumab, we still had not reached the median survival. But with that being said, the hazard ratio was 0.58, and this was highly significant. So a 42% decrease in the risk of death with the combination. This is the first time in over a decade that we've beaten the survival with a frontline TKI like Serafinab. And this data was supported by the improvement of PFS with a hazard ratio of 0.59. Again, a first with a novel regimen over Serafinab. Now, if we look at the independent review by response, you can see that 27% of the patients in the combination arm had a confirmed response with about 5% of those being complete responses. And we still had not reached the median duration of response at the time of this review. This was the same pattern seen with modified resist with a 33% response rate. Now, looking at adverse events, overall, they were very comparable. 98% of patients in each arm had some adverse event. Grade three, four events were very similar between both arms. There were actually more grade five events in Serafinab arm, 6% versus 4.5%. And when we look at SAEs or serious adverse events, 38% with atezobev versus about 31% with serafinib. Now, there were more adverse events leading to withdrawal of the drugs, either drug, but still relatively low, 15% for this combination versus 10% with serafinib. And this definitely compares favorably to other TKIs in phase three studies. These are the most common adverse events, and you can see that when we use the combination, we're not seeing a synergistic effect. We see adverse events from atezolizumab, which are generally autoimmune or infusion reactions, and those from bevacizumab, which are from its anti-VEGF activity, which is really hypertension and proteinuria. And certainly this can occur with serafinib, but higher grade with atezobev. There were some enzyme abnormalities as well occurring in both arms, slightly higher with atezobev. Now, looking at one of these graphs in lower color, lighter colors, all grade adverse events, and darker colors are grade three, four events. And again, a similar pattern, but keeping in mind those adverse events that patients notice the most, diarrhea, hand, foot, skin syndrome, occurring more commonly with serafinib. And the same pattern for treatment-related adverse events. Again, hypertension and proteinuria with atezobev being most common, hand, foot, skin syndrome, and diarrhea with serafinib. Now, serious adverse events. We can see here that there was about, just over 1% of patients had a grade five bleeding event with atezobev, and none were seen with serafinib. But needless to say that when selecting patients carefully, we can minimize the likelihood of these events. Now, there's other combinations being looked at. Linvantinib, the VEGF multikinase inhibitor in combination with pembrolizumab has shown response rates of 36% in a single-arm phase two study. Very impressive data with duration of response over a year and a disease control rate of 88%. This is now being looked at in a phase three study of Len and Pembro versus Len alone. This is a blinded randomized phase three study, and we're waiting for the readout. There's also interest in dual immunotherapy. Ipilimumab and nivolumab was evaluated in several dosing regimens in the second line setting, consistently giving response rates of over 30% with a long duration of response. And actually, this arm A had a median survival of over 20 months. And based on this, this regimen got accelerated approval in the second line setting, though keep in mind about 50% of patients needed steroids to manage the autoimmune adverse events. This now is in a phase three study of ipinivo versus serafinib in the frontline setting. And dervalumab, a PD-L1 antibody in combination with tremulumab is also being looked at. And in this study presented at ASCO this year, we saw a overall response rate of about 18% confirmed, 25% unconfirmed. And again, a fairly well tolerated regimen, less adverse events than we saw with ipinivo. And this is being evaluated in the Himalaya study of the combination versus dervalumab or serafinib and events are weighted. And also finally, the COSMIC study is looking at cabizantinib, multi-kinase inhibitor approved in second line with atezolizumab versus serafinib. This study is ongoing as well. We have not really seen any early phase data with this combination, but needless to say, I think they are betting on consistency with the linvantinib data set and the bevacizumab data set. So we now have many drugs approved in advanced liver cancer and how we sequence these is still yet to be determined. Certainly based on the efficacy and safety data, atezobev would be considered frontline. And then there's various opinions on sequencing the TKIs and ramicerumab and other drugs, but there's certainly an interest of moving these new novel combinations and many phase three studies are looking at them in earlier stage liver cancer in combination with case or even adjuvantly after surgery or RFA. So in conclusion, we have many positive phase three studies with TKIs and ramicerumab, but now we have the first superiority study in frontline with atezobev. Again, we need to figure out how best to sequence these drugs and there's many ongoing studies, which I look forward to seeing the results of, of TKIs and immunotherapy as well as immunotherapy combinations in advanced and earlier stage of liver cancer. Thank you very much for your attention. I hope you found this useful and as always, I appreciate you inviting me and hope everybody's well and hopefully we'll be together soon at meetings for the end of COVID. Thank you very much. Hello, it's my pleasure to participate in a symposia by the Palo Verde Neoplasia Interest Group. I thank the organizers for this invitation. I am a hepatologist at the Mayo Clinic in Rochester, Minnesota. I have a longstanding interest in clandrocarcinoma including basic and clinical research. I've co-authored multiple publications on this subject. Today, I'll be speaking about advances in precision oncology in intrapathic clandrocarcinoma. I have no disclosures relevant to this presentation. Clandrocarcinoma may be divided by anatomic subtypes and growth patterns. Intrapathic clandrocarcinoma is mass-forming lesions in the hepatic parenchyma, which will be the subject of today's seminar, but we should not forget that there's also perihilar clandrocarcinoma involved in the hilum of the liver and distal clandrocarcinoma involving the distal bile duct. I will refer to as intrapathic clandrocarcinoma as ICCA for the remainder of this talk. Not only are there differences in the anatomic location of these different subtypes of clandrocarcinoma for clinical presentation and treatment, but there are also differences in their genetic alterations. And we have made the major advances in regards to defining the genetic alterations which are targetable in clandrocarcinoma in the ICAA, or the ICCA subset. Three years ago, Dr. Rizvi and I suggested an evolving roadmap for the molecular stratification of clandrocarcinoma with therapeutic implications. Here we suggested that clandrocarcinoma tissue could be subjected to diagnostic assays such as fluorescent in situ hybridization, DNA or RNA sequencing, or immunohistochemistry, and subdivided into different subtypes of clandrocarcinoma, which could then be targetable. Today we'll show you what we have learned in the last three years. First, I'll speak about genetic fusions. Genetic fusions are common in clandrocarcinoma, especially those of the FGFR, or fibroblastic olfactory receptor 2 subtype, and also in neurotrophin receptor tyrosine kinase genes. Indeed, fibroblastic olfactory receptor, or FGFR, has four types of receptors, one through four, and all four subtypes have been implicated in human cancer, and thus this family of receptors is indeed autogenic in humans. And note that the FGFR2 fusions may occur in up to 45% of intrapiglandrocarcinomas, but a more common number would be 15 to 20%. And this family of genes can also be amplified and dysregulated in the paloseric carcinoma as well. The FGFR2 fusion abnormalities can be therapeutically targeted. This summer, in NASA Oncology, we had the first report of pemegatinib, employed for previously treated, locally advanced, or metastatic clandrocarcinoma in a multi-center open-label phase two trial. 146 patients, including patients with fusions, other genetic alterations, and those with no alterations. The primary endpoint was proportion of patients with objective response, and the results in those with fusions was impressive enough to warrant approval by the FDA, and this was indeed the first drug ever approved for treatment of clandrocarcinoma. First endpoint was the best percentage change from baseline in target lesion size for individual patients with FGFR2 fusions or rearrangements. Those without fusions or rearrangements had no response, and most of the dramatic responses occurred in those with fusion abnormalities, and as you can see by this waterfall plot. The overall survival, or progression-free survival, as shown here, was about six to nine months in this heavily treated patient population. This was impressive enough when combined with the waterfall plot to lead to approval of this drug following standard of care therapy for clandrocarcinoma. We should be aware that drugs like pemegatinib are reversible inhibitors of the FGFR2 kinase pocket, that mutations can occur, resulting in drug resistance and disease progression. And that there are irreversible inhibitors of the kinase target being tested in human trials, and one of these drugs is TAZ120. The trials are ongoing, and the results have yet to be reported, but they likely will be more promising than those with the reversible inhibitors. The NTRK pathway is also a classic tyrosine kinase pathway, and fusions of this receptor may also occur leading to its constitutive activation. NTRK fusions are quite common in human cancers, both in adults and pediatrics, and particularly in high-grade gliomas, but also in clandrocarcinoma. And in a basket trial published in the New England Journal of Medicine, merotrectinib was employed for patients who had NTRK fusion abnormalities independent of tissue of origin. This is termed the basket trial, and you can see the waterfall plot for this patient population. And again, a few patients with clandrocarcinoma were included and had responses. I will now leave the genetic fusions and focus on genetic alterations. There are two I'd like to discuss, one being disocitrate dehydrogenase mutations, which are gain-of-function mutations, and the BRAP600E, which are also gain-of-function mutations of the kinase pocket. There's been a lot of interest in isocitrate dehydrogenase mutations. First, it should be noted that these have been described in brain tumors, leukemias, and a rare subtype of sarcoma, but really are only prominent in one type of GI cancer, that clandrocarcinoma. They're not common in gastric, colorectal, pancreatic, or esophageal cancer. This mutation is interesting. So the wild-type enzyme, which is present in the Krebs cycle results in the generation of alpha-Q grulorate, which then is metabolized to succinate as a substrate for the mitochondrial electron transport pathway. Those patients have mutations of IDH1 or 2, metabolized alpha-Q grulorate to 2-hydroxyglulorate. 2-hydroxyglulorate accumulates in the cell and interferes with those enzymes using alpha-Q grulorate as a substrate so that 2-hydroxyglulorate can be considered oncometabolite. The IDH mutations are common in intrapathic clandrocarcinoma, occurring in approximately 25 to 30% of patients. The alpha-Q 2-hydroxyglulorate can be measured in plasma and it may be a pharmacodynamic marker for the drugs being employed, and a number of drugs have been developed to treat patients with IDH1 or 2 mutations. I'd like to review the data with ivositinib and IDH1 mutant chemotherapy refractory clandrocarcinoma. This was a multi-center, randomized, double-blind, placebo-controlled phase three trial. It's a 2-to-1 randomization of active drug to placebo in 185 patients. The drug was, the study was confounded by crossover design in that patients who progressed on placebo then crossed over to active drug. The primary endpoint was progression-free survival on intent-to-treat analysis. Progression-free survival was assessed by independent radiology center before crossover and intent-to-treat analysis. You can see that there was a clear separation between those on active drug versus placebo in regards to progression-free survival. Overall survival was confounded by the crossover design. This green line reflects an attempt to correct for this mathematically, but again, this is a mathematical correction. You can see that the difference at six months was apparent in terms of overall survival on those on drug versus placebo. The FDA is still assessing these data for potential approval of ivo-Citinib for the treatment of patients with chondrocarcinoma who have IDH1 mutations. The last mutation I'd like to review is the BRAF V600E mutations. They are mutated. This is a L-cell gain-of-function mutation in a subset of patients with biliary tract cancer. And I will review the results of the ROAR trial, which was, again, a basket trial, a tissue-agnostic trial between patients who had BRAF V600E mutations. Based on the lessons learned in melanoma, where BRAF mutations are common, a MEK inhibition is required to overcome compensatory survival signaling pathways. And so, a BRAF inhibitor plus a MEK inhibitor was employed. Overall survival was 11.5 months, equivalent to those from a historical cohort treated with gibbine-cytobine cisplatinum, the current standard of care. This was a waterfall plot based on attention-to-treat analysis. Again, you can see that the majority of patients had either a partial response of stable disease or progressive disease, so they had a high disease control rate. And overall survival at about one year was, again, approximately 50%. And so, this mutation can be targeted and treated, much like it can be in melanoma in those patients with cholangiocarcinoma and BRAF V600E mutations. Finally, I'd like to examine what we know about immunotherapy and cholangiocarcinoma. Note that cholangiocarcinoma is a highly desmoplastic cancer with a very rich stroma. One can see the glands present here, but also note the rich stroma of both matrix and immune cells, including NK cells, NKT cells, tumor-associated macrophages, cancer-associated fibroblasts, and T cells. And indeed, a great deal has been learned in experimental systems regarding the role of cancer-associated fibroblasts as mediators of disease progression in these cancers. These cancer-associated fibroblasts speak to and modulate the activity of a wide variety of immune cells, as illustrated in this slide. Unfortunately, we do not yet have a clinically viable approach for targeting cancer-associated fibroblasts for the treatment of this cancer, although this could maybe be a very potentially useful strategy. More importantly, we're all aware of the role of immunotherapy targeting checkpoint inhibitors in tumor immunotherapy. One can target either the C2A4 pathway with monoclonal antibodies or the PD-1, PD-L1 pathway with antibodies to overcome T cell exhaustion and allow cytotoxic T lymphocytes to eradicate the tumor. We know that this approach is extremely effective in patients with DNA mismatch repair. They show the waterfall plot of science, again, in a basket trial, which was tissue agnostic based on the presence of mismatch repair deficiency, and patients with chlangic carcinoma, as well as others, had good responses. Unfortunately, most patients with chlangic carcinoma are DNA repair proficient. But this is not discouraged. A number of trials looking at checkpoint inhibitors targeting the PD-L1, PD-1 pathway or the C2A4 pathway. This is a result of a study published in JAMA Oncology this summer. Patients with biliary tract cancers, including gallbladder, ICCAs, and patients with distal and perihelic chlangic carcinoma classified as extrahepatic here. They noted responses in the patients with gallbladder cancer and intrahepatic chlangic carcinoma in an entire cohort. The one-year survival was approximately 50%. A similar study targeted the PD-L1 pathway for biliary tract cancer using a phase two single-arm study. Again, it was a mixed study of patients with biliary tract cancer, including gallbladder, intrahepatic, and patients with extrahepatic chlangic carcinoma. They observed responses in all subgroups in this study, and again, around a 50% 12-month survival. But this is an evolving field, and there are a wide variety of approaches to targeting the immune environment in these desmoplastic cancers. This includes ways to look at enhancers of NK cell activity, promote targeting tumor-associated macrophages in combination with checkpoint inhibitors, as we've reviewed. But there are also non-T cell approaches to addressing cancer by immunotherapy approaches, including approaches to inhibit myeloid drive suppressor cells, for example, using CAR-T cells, using cancer vaccines, and a wide variety of imaginative approaches that have promise in cancer immuno-oncology. So where are we now? Three years later, we talked about this stratification approach to intra-hepatic cholangiocarcinoma. I think we've shown, I think we've learned that IDH mutations are targetable to the benefit of patients, as are patients with FGFR2 aberrations, and then patients with DNA mismatch repair deficiency, immune checkpoint behaviors are efficacious. We have a lot of work to go, as even in IDH patients and patients with FGFR2 fusion amplification, the majority of patients have primary resistance, and the disease response is not durable in regards to the long-term effects, and so combination therapy will be necessary. And then for patients who have DNA repair proficiency, which is the vast majority of patients with cholangiocarcinoma, a great deal of work needs to be learned using immuno-oncology. So the takeaway points is that intra-hepatic cholangiocarcinoma is a genetically heterogeneous disease with several identifiable driver mutations, and the major ones targetable to date are the FGFR2 fusions, the IDH mutations, but we should not forget the BRAF mutations, or the NTRK fusions. Hence, patients with intra-hepatic cholangiocarcinoma should have their specimen sequenced to identify targetable mutations because of the therapeutic implications. The role of immunotherapy for treatment of this disease is an area of intense investigation ongoing. So thank you. Hello, my name is Katie Kelly, and I'm a GI medical oncologist at the University of California, San Francisco. I'd like to thank the organizing committee for the honor of speaking today on the role of molecular tumor boards for the management of liver cancer, specifically HCC and intra-hepatic cholangiocarcinoma, or ICCA. Here's a brief biography, and here are my disclosures. Here's the outline for the talk today. The evolving role of molecular tumor boards in oncology will be the starting point. I'll then review key therapeutic targets in primary liver cancers, specifically hepatocephalic carcinoma, or HCC, and intra-hepatic cholangiocarcinoma, or ICCA, and then discuss in further detail the potential roles for molecular tumor boards in hepatocephalic carcinoma and ICCA with some case examples and some important caveats and challenges, followed by key takeaways. So starting first with molecular tumor boards in oncology, or MTB. MTB have emerged in most cancer centers in response to the rapid proliferation of genomically targeted therapies in oncology. The goals of MTB are really primarily to focus on treatment decisions according to tumor mutation profile and guide the choice of tumor-specific FDA-approved therapies, tumor-agnostic FDA-approved therapies, decision-making about using off-label therapy according to mutations in one tumor type that have evidence in another tumor type, or allocation to clinical trials based on mutation profile. MTBs can also play a role in guiding or helping to clarify diagnosis in the case of an uncertain primary cancer and identifying germline or hereditary mutations. And to the point of tumor-agnostic FDA-approved therapies, I just wanted to show here one of the classic examples of why the field has evolved in this way. Drugs like larotrectinib, this is a targeted inhibitor of activating intrec fusions, which are present across a variety of tumor types, and in large basket trials for larotrectinib as well as a similar drug, intrectinib, we see really dramatic tumor responses irrespective of histology or tumor site of origin, including a cohort of patients with complete responses that have lasted for quite a long time. And this kind of efficacy is why we practice molecular medicine in oncology, certainly the holy grail of what we do, and also has led to tumor-agnostic FDA-approved drugs such as larotrectinib, which are approved for use and warrant testing in any solid tumor that is known to have potential for intrec fusion. So more on molecular tumor boards in oncology. The membership of these molecular tumor boards includes generally pathologists, geneticists, or genetic counselors, noting again the potential to uncover germline mutations in this process of tumor genotyping, as well as oncologists. The MTV generally will perform review of NGS results across a variety of tumor types in any given tumor board, but it's important to note that case presentation, of course, requires the availability of tumor tissue or circulating tumor DNA or ctDNA to be tested for next-generation DNA sequencing operations. So here are two relatively large academic centers' tumor profiling results associated with molecular tumor boards at the Memorial Sloan Kettering impact cohort of over 10,000 cases, and here from UCSD Cancer Center, a smaller cohort of about 715 cases. I wanted to highlight here that biliary cancer, such as intramenoclangioplasm, is relatively prevalent in both of these cohorts, 5% of this molecular tumor board cohort and about 2% of the impact cohort, but HCC, much lower for both of those, even though it's a vastly more prevalent cancer, and we must—this is probably for multifold reasons, but one of which is the lack of available tumor tissue for testing in many cases. And another point about molecular tumor boards is that the utility of such a tumor board is, of course, dependent on both the presence of an actionable mutation as well as the detection thereof. This bears backtracking a little to define actionability, and this is a bit of a fuzzy definition, but one accepted standard is this actionability table that ranks actionability according to whether it's FDA-recognized in the same indication, whether it's a biomarker used outside of FDA-labeled in the same indication, and then whether it's crossing into other indications and based on lower levels of evidence. So, if we look here back to that Memorial Sloan Kettering impact cohort of over 10,000 cases, and if we look at the presence of actionable mutations, and based on these definitions in biliary cancer, we see that almost 50 percent have some actionable mutation, whereas much fewer in HTC, less than 10 percent. And that brings me to the next section, what are the established therapeutic targets in HTC? This is a very nice schema from a Nature review paper, which shows FDA-approved drugs at the time in green, those that have been negative in orange, and those in study in yellow. It highlights the plethora of potential or candidate therapeutic targets ranging from receptor tyrosine kinases, intracellular kinases, immune response, cell cycle, and genetics, and brings us to the point that none of these targets have a known genomic biomarker predictive of response. And why is that? If we look at large sequencing studies of HTC, such as the TCGA, we can see that really the majority of mutations, the most prevalent mutations, do not have…are not conventional drivers that have a drug that inhibits them, and so most of these do not meet the criteria of actionable to date. And in fact, if we look at the drugs that are approved by the FDA for HTC, which obviously is a very rapidly expanding landscape over the past few years, none of them have genomic biomarkers predictive of benefit. Now, gramiserumab benefit is predicted by high serum alpha-fetoprotein, but that is not a genomic marker. And one of the challenges, to return to the theme of limited tumor tissue, is that when we look at registrational trials, from the SHARP trial of SRAF-LIB to the more recent INBRADE-150 trial of teslozumab plus ebicismab, only a paucity of patients in all of these trials have had adequate tumor tissue for biomarker testing, let alone next-generation sequencing. So not only are the genetics of HTC very difficult to target, but we also are not collecting that tumor tissue to study them further. So are there any candidate genomic biomarkers in HTC? Well, there are some. Turning to the tumor agnostic genomic biomarkers predictive of treatment response across a variety of tumor types, things like the Ntrek fusion that I mentioned on the first slide, those are quite rare and actually have not been reported in HTC to date. In the FDA-filed studies, emicrocyte high, emicrohexylate instability high, tumor mutation burden high are tumor agnostic indications for immunotherapy of pembrolizumab, but these are quite rare in HTC as well. I think there are some actual mutations that we are certainly interested in, including MET amplification in a subset of patients. So a phase III trial of MET inhibitor in MET high by protein expression was negative using the drug Tivantinib. FGF19 amplification, that's the ligand for FGFR4, the receptor tyrosine kinase. The FGFR4 inhibitor, Fizogatinib or Blu554, did show as a proof of principle tumor responses in about 17% of HTC patients with FGF19 over expressed HTC, so this is an interesting target to look at as a potential future genomic biomarker. And there are others being studied, but nothing else with strong prospective data to date. Now, the story is quite different for intrapartic cholangiocarcinoma. We now have a variety of established genetic targets, including some with biomarkers, and I'll focus on a couple of these here. FGFR2 fusions, which are receptor tyrosine kinases with fusion mutations present in about 15% of intrapartic cholangiocarcinoma. We talked already about Ntrek fusions that are very rare but have been reported in cholangiocarcinoma unlike HTC. Intracellular kinases like BRAPV600E mutation present in 5 to 7%. IDH1 mutation, a tumor metabolism operation in 13% of intrapartic cholangiocarcinoma. And microcyte high or tumor mutation burden high, which correlate with immune responsiveness to immunotherapy with pembrolizumab or other PD-1 inhibitors present in 2 to 3%. So in sum, if you collectively group all of these, we see that there are tumor genomic biomarkers predictive of response in over 30% of intrapartic cholangiocarcinoma patients, even conservatively speaking. So taking these one at a time, I'll first talk about FGFR2 fusions. As I mentioned, about 15% of intrapartic cholangiocarcinomas have these. They are intronic translocations that can have many different fusion partners, and they constitutively activate FGFR2 signaling. And we've seen in a host of drugs now, three shown here, Dugratinib, Pembigatinib, and Fudubatinib, that FGFR2 inhibitors elicit very prolonged and durable strong responses in these FGFR2 fusion cholangiocarcinoma patients, and that led to FDA approval of Pembigatinib earlier this year. And there are multiple other agents in development, and we may see that their mechanism of action and target profile may enable sequential therapy to overcome resistance. Just a brief note on that, this really interesting rapid serobiopsy plasma cell-free DNA testing and rapid autopsy study from Goyal et al. in Harvard showed that the mechanism of resistance to FGFR2 inhibitors is the development of kinase-like domain mutations in the coding region of FGFR2. Which themselves may be susceptible to different FGFR2 inhibitors down the road as a potential role for sequential therapy. Another target in intrapartic cholangiocarcinoma is the BRAFB600E mutation, which is a known target in melanoma and colon cancer, but recently was shown in the basket trial called ROAR. Also to be a potential predictive biomarker, a genomic biomarker of response with the combination of debrafinib, a BRAF inhibitor, plus trimetinib, a MEK inhibitor in cholangiocarcinoma, based on the finding of very durable, prolonged, deep responses in a treatment refractory population. Only 43 patients, but this is a phase II trial. And again, the fact that BRAFB600E is an established biomarker in other tumor types, and that we see a very similar result now in cholangiocarcinoma with the BRAF and MEK inhibitor combination, I think does support considering off-label use of these drugs, which are approved in melanoma and other cancers, for cholangiocarcinoma as well in the setting of treatment refractory disease. IDH1 mutations are a really interesting type of mutation that lead to differentiation block due to the accumulation of the downstream alkylmetallite to hydroxyglutarate. An inhibitor of IDH1 mutant, IDH1 ibucitinib, or AG120, was studied in intrapartic cholangiocarcinomas, which have about 13 percent of those have IDH1 mutations, I should add, in a phase III trial called the CLARITY trial. The primary endpoint was progression-free survival, and you can see here a dramatic separating of the treatment arm with ibucitinib compared to placebo, hazard ratio of 0.37, which was strongly significant and has led to the inclusion of ibucitinib in the NCCM guidelines for intrapartic cholangiocarcinomas harboring IDH1 mutations. Lastly, speaking of intrapartic cholangiocarcinoma, I think it's always important to touch upon the role of testing for mismatch repair deficiency and or microsatellite instability, mismatch repair deficiency being the protein deficiency and the microsatellite instability being the genomic signature or phenotype of mismatch repair deficiency in a tumor. About 2 to 3 percent of cholangiocarcinomas are MSI-high or deficient in mismatch repair, and we know from basket trials across tumor types that patients with MSI-high or DMMR tumors are extremely sensitive to PD-1 inhibitors such as pembrolizumab. This was a basket trial which, again, shows high-rate objective response around 20 percent of complete responses, and this included patients with cholangiocarcinoma. This led to a tumor agnostic approval of pembrolizumab for patients with these mutations and warrants MSI and DMMR testing in cholangiocarcinoma. Returning to a few cases, here's one case, a 57-year-old female with no underlying liver disease. She developed abdominal pain and fullness, was found to have a Lyrax V liver lesion with metastases and portal vein tumor thrombus, IVC invasion, a biopsy because, again, she had no known liver disease she did undergo at biopsy. It showed poorly differentiated HTC, and this biopsy was sent for tumor next-generation sequencing, which showed a very high copy number metamplification, which is reported in a subset of around 6 percent of HTCs. She was originally treated with serafinib followed by a PD-L1 inhibitor, but unfortunately in rapid progression, and then as her third line therapy, cabozantinib was chosen because of the high copy number metamplification, knowing that cabozantinib, the receptor tyrosine and kinase inhibitor, does have inhibition of MET, and she then had a striking AFP response and tumor stabilization prolonged for over 11 months, which was much longer than she had with either of her first and second line therapies. So in this case, her tumor sequencing and potential molecular tumor board involvement guided the choice of a standard therapy based on her tumor next-generation sequencing results. Here's another case of a 61-year-old female with NAFLD and obesity who developed abdominal pain. A CT scan showed an infiltrated tumor in an expansile right portal vein tumor thrombus. She was diagnosed radiographically as HTC at an outside hospital and treated with Y90 and TACE. A repeat scan, though, showed worsening biliary obstruction requiring a stent. She also had progressive peripheral enhancing centrally hypodense lesions and some ascites, and this radiographic appearance prompted a biopsy, being atypical for HTC at that point, and she was found to have a poorly differentiated carcinoma with some glandular features. We then sent tumor next-generation sequencing, which showed FGF1-2-BF1 fusion, which is characteristic of intrapathic cholangiocarcinoma and not at all of HTC, very unlikely or rare in HTC. So she was started on chemotherapy for glandular carcinoma with the GEMSYS regimen, but unfortunately progressed quickly. We then treated her with TAZ120, which is also known as Pudibatinib, an FGFR2 inhibitor, a covalent inhibitor, on a Phase II trial, and she experienced a prolonged partial response lasting over eight months. Again, highlighting the importance of tumor next-generation sequencing and potentially longer tumor board involvement, not only to clarify her diagnosis in this case, but also allocate to an effective treatment. This was on a clinical trial. Lastly, another case, a 74-year-old Asian male with chronic active hepatitis B who was appropriately suppressed on entechamere, had a left lobe lesion on screen, which was resected and found to be a poorly differentiated adenocarcinoma compatible with intrapathic cholangioma. He unfortunately had recurrence within about six months in the port of hepatitis and retroperitoneum. Here's the lymph node, and a repeat biopsy with tumor next-generation sequencing showed a P53R249S mutation consistent with prior aflatoxin exposure. He was of note born in China, but also a high copy number EGFR amplification, thought to be a driver. He was treated first with gemcitabate, cisplatin, and an EGFR monoclonal antibody based on data for using a patent tumor out of the EGFR monoclonal and high amplification EGFR mutation tumors. And then he actually did quite well on that for over a year, but when he progressed, he was found to have a new EGFR mutation, G719A, in addition to the amplification. So we reviewed his case at the Moncure Tumor Board and elected to treat him with the drug efatinib, which is FDA approved for lung cancer, off-label, based on the efficacy of this drug in EGFR G, I should say, 719A exon 18 mutated non-small cell lung cancers, but also in other solid tumor types. And here's the phase two data, which shows that efatinib does have responses in other tumor types, but not beyond lung cancer. And his treatment is ongoing. So some important caveats and challenges before concluding. Of course, Moncure Tumor Board requires access to tumor sample with sufficient tissue for NGS and the presence of actionable findings. Circulating tumor DNA axes can be used for certain exonic or hotspot mutations in the coding region, but generally don't cover intronic regions, and so can't be used for fusion detection, which are usually intronic. I think I'll skip to here to comment that the impact of the mutation or targeted therapy can be context dependent, meaning the efficacy in one tumor type may not extend to another tumor type. In the classic example, this is BRAF inhibitors do not work well as monotherapy in colon cancer, though they do work extremely well in melanoma. So it's important before using off-label therapies to check if there's been at least some early phase basket, phase two trial data to infer whether an agent may have cross-cutting efficacy in a different tumor type. And lastly, very important consideration to be aware of the risks of off-label therapy, the safety of drugs in one tumor type may not have been studied in the context, for example, in the case of HTC or glandular carcinoma in the context of hepatic dysfunction. So this requires careful checking of the label for dosing and liver dysfunction in serotic patients. There can be financial risk or distress if a patient does not get approved for the drug or has a high co-pay, and there can be a lot of provider and staff workload to appeal or achieve expanded access program access for off-label drugs. So key takeaways. Tumor next-generation sequencing should be performed in all advanced intrapathic glandular carcinoma patients. This is for FGFR2 and intrafusions, BRAF, and IDH1 mutations, and MSI-high or mismatch repair deficiency, which predict response to targeted or immune therapies, respectively. I think we should be considering tumor NGS in HTC patients who have atypical diagnostic features or equivocal diagnosis or who may be eligible for clinical trials, but remembering that there are no currently established tumor genomic biomarkers for treatment response to HTC to date. And I should also redouble our efforts to collect tumor samples in prospective clinical trials for this reason and focus on biomarker research. And lastly, importantly, with the advent of tumor next-generation sequencing in liver cancers, molecular tumor board review can help us guide the choice of therapy for these actionable mutations, help clarify the diagnosis at hand when equivocal diagnoses emerge, and help us identify patients who may require germline testing. So with that, I'll conclude by saying thank you for your attention and refer again.

Liver cancer

Hepatocellular carcinoma

Intrahepatic cholangiocarcinoma

Trial design

Immunotherapy

Precision oncology

Management strategies

Immune checkpoint inhibitors

Anti-angiogenic agents

FGFR2 fusions

BRAF V600E mutation

Tumor sequencing