Mr. Chairman, organizers, and colleagues, it's my pleasure to present a lecture on precision medicine to guide therapy in liver cancer. Here you have my disclosures. So as has been mentioned, we have a case of a 61-year-old man with alcoholic cirrhosis on routine MRI has found to have an unusual two centimeters mass in hepatic segment 6 with a thick viable enhancing and restricting rim. So as has been described nicely before, this represents a Lirat's M case, and according to the guidelines of SLD, these cases are recommended for a multidisciplinary discussion and for biopsy. So there are several studies analyzing what is the final diagnosis of Lirat's M, and in this study, the authors nicely analyzed more than 2,700 cases, and among those that were classified as Lirat's M, 36% were in reality HCC, but 93% were actually malignant. This is another series of up to 1,000 lesions in cirrhotic patients, and among those again classified as Lirat's M by histology, 44% of them were HCC, but 56% were non-HCC malignant as well. Therefore, in this case, the recommendation, of course, is to biopsy the patient, and what type of information can we obtain from this biopsy. Of course, the diagnosis on one hand, we can define also biology, molecular classes, the prognosis, and eventually, I will try to provide a frame of what can we do in terms of precision oncology, both for HCC and cholangiocarcinoma. As per guidelines of ASLD and ESL for HCC and guidelines of ILCA for intrapartic cholangiocarcinoma, the pathological diagnosis of HCC should be based on international consensus recommendations using required histological, immunohistological analysis, and these guidelines, in case of differential diagnosis between hepatocellular carcinoma, cholangiocarcinoma, mixed tumors, and immunostaining with glypicon-3, heteroprotein-17, and glutamine synthase. Therefore, in this case, after using all these methods, we can have a differential diagnosis of atypical HCC, probably less likely, intrapartic cholangiocarcinoma, or even other malignancies. Let's see how we can obtain from the biopsy in terms of biology and molecular classes. The molecular classification of HCC has been established in at least two subgroups, the proliferation class and the nonproliferation class. Within the proliferation class, we recognize what is called S2, or a class with progenitor cell features, and generally is associated with poor outcome. There is an S1 subclass driven by TGF-beta signaling. In the nonproliferation class, the more characterized subgroup is enriched by beta-catenin mutations and Wnt signaling activation. These classes have been correlated with clinical and pathological features, and relative proliferation class is associated mostly with patients with hepatitis B virus infection, poorly differentiated tumors, high FP levels, particularly the progenitor cell class, and worse outcome. As opposed to the beta-catenin subclass, in which is mostly associated with alcohol-related HCC-beta-DC virus infection, and better outcome. In terms of linking these with the actual treatment, we only have evidence that patients with high AFP that represent the S2 subclass, or what is called the progenitor cell subclass, these patients are responding to RAMU0MAP in advanced cases. We also may be able to define the immune classification of HCC. We reported this study a couple of years ago, 24% of the tumors actually are classified as immune-hot, with enrichment of T cells, CD8s, NK cells, TLS, and cytotoxic cells, and also gene signatures related with response to checkpoint inhibitors. In reality, this represents between 20% and 30% of the cases, and on the other hand, we have 25% of the cases of tumors that we call immune-excluded, with a paucity of T cells or B cells, and enrichment of beta-catenin mutations. This has not been proved yet, and is still under testing, but the hypothesis here will be that the patients which tumors have an immune-active class may be responders to checkpoint inhibitors. Also, cholangiocarcinoma has been classified according to molecular traits, and this classification leads to two subclasses, the inflammation class with better outcome, in this case 47 months median survival after resection, as opposed to the proliferation class, 23 months median survival after resection. And here you have detailed all the molecular features comparing the inflammation class and the proliferation class. In terms of prognosis, also we can obtain information from the biopsy. We have to understand, however, that the outcome of HCC is not only based on the tumor biology, but also based on the non-tumor serotic features, as we published years ago in the New England Journal. So the information from the tumor is not enough to predict survival of these patients. I have summarized in this table the gene signatures and biomarkers that actually have prognostic power in HCC. We have signatures from the tumor, the 5-gene signature, the Epcan signature. We have signatures from the adjacent tissue that also have prognostic power in terms of survival. This will be the immune-based signature, the PUN survival signature. And finally, we have prognostic markers. And as you can see, the prognostic marker of AFP more than 400 nanograms milliliter is the only one actually adopted by guidelines as having prognostic power, and in fact, this threshold is currently used to stratify patients in clinical trials. The most exciting area, however, is how we can use the biopsy to treat the patients, what is called precision oncology. Let's start with the treatment strategy of HCC, according to SLD guidelines, or the HL guidelines. This case, and Dr. Kollig will talk deeply on that, and refers to very early tumors that probably will be treated with these strategies. But at the advanced stages of the disease, the management is based on systemic therapies. Systemic therapies that are currently approved by guidelines are as follows. First, we have in front line, sorafenib and lembatinib, and in second line, regorafenib, cabozantinib, and ramucirumab. These five treatments have been shown to expand the survival of the patients. Also, we have nivolumab and pembrolizumab that have been approved by FDA based on phase two data. However, a recent trial comparing nivolumab versus sorafenib in front line show a non-significant trend favoring nivolumab, where it has a ratio of 0.85, but the trial was negative. Interestingly enough, though, the patients receiving nivolumab, 15% of those achieved objective response rate, and this was a long-lasting objective response. Same happens with pembrolizumab versus placebo in second line. These patients were randomized to receive pembrolizumab or placebo after front line therapy, and as you can see, there is a hazard ratio favoring pembrolizumab of 0.78. That was not enough to hit a positive survival endpoint. Nonetheless, patients receiving pembrolizumab, again, achieved 18% of them objective response and long-lasting responses. So therefore, the search for biomarkers predicting response to checkpoint inhibitors in HCC is a clear unmet medical need, and therefore, a biopsy can help in that. So which are the biomarkers that in oncology have helped so far to identify patients responding to checkpoint inhibitors? Out of this thorough list, only two biomarkers have been validated in phase three trials. First, tumor mutational burden, particularly in melanoma and non-small cell lung cancer, and PD-L1 expression, particularly in non-small cell lung cancer. In HCC, as I have mentioned, we have the immune class, immune-hot tumors on one side, and we also have the immune-excluded class, characterized by enrichment of beta-catenin mutations. The assumption is that these patients of immune-excluded class eventually may be those that have primary resistance to checkpoint inhibitors, and in fact, we recently published this year a study in experimental models of HCC that suggests that beta-catenin activation promotes immune escape and resistance to anti-PD-1 in hepatocellular carcinoma. In these lines, a recent study from the group of Memorial, including only 27 patients, it's just a pilot study receiving checkpoint inhibitors, have identified that those patients that are resistant to checkpoint inhibitors actually have an enrichment of beta-catenin or axin-1 mutations and certainly a worse prognosis. So this gives us a frame of the potential use of biopsy in identifying responders to checkpoint inhibitors. This situation has certainly changed very recently. Just a few weeks ago, there was announced a positive randomized control trial with a combination of a tesolizumab plus bevacizumab versus sorafenib in terms of overall survival and progression-free survival. So the study was stopped for superiority of this combination at the interim analysis, and the results will be released at the end of this month. Also from the biopsy in oncology, a lot of treatments have been approved. This is the case, for instance, in breast cancer, of amplification of Ertunol in 25% of the cases and approval of trastuzumab, BRAF mutations in melanoma, 60% approval of bemoraphenib, or even ALK fusions in 3% of small cell lung cancer, approval of clisotinib. All these approvals have been based on biopsies. In reality, in this study of somehow real life, including 10,000 patients, all of them with tumors in which the genes and the potential rivals have been screened with the IMPAQ platform assessing 500 genes. Here you have the list of potential drivers that have been identifying these 10,000 patients. Actually the authors identified 37% of the patients with actionable targets that actually receive precision oncology based on the description of these mutations. Well where is the situation of HCC regarding this? Well in reality, here you have a meta-analysis of the most relevant mutations in HCC, and the most frequent mutations, third mutation, P53, beta-catenin added one, are unactionable, undruggable, as opposed to the less frequent mutations that are actionable. In this meta-analysis we're showing that 25% of the patients have at least one actionable mutation. This means that 75% of the patients do not have any actionable mutations. And in the context of oncology, here you have listed the major cancers, but the cellular carcinoma, as you can see, has 25% of actionable mutations, so it's one of the tumors with less actionable mutations, and for sure there is no FDA approved drug based on that. There have been some attempts to change that, and here I'm showing the data of FGF19. FGF19 have been shown to be an oncogene in experimental models. It's amplified in 7% of HCCs, and overexpressed in 20% of HCCs, and with this drug, BLU554, that is a very selective drug blocking FGF receptor 4, it has a lot of activity in PDX models, as you can see here. As a result of this preclinical study, there was a design of proof of concept with patients with HCC in second line with FGF19 positive bi-immunostaining, as you can see here, compared with FGF19 negative bi-immunostaining. And certainly those patients with FGF19 positive treated with fisogatinib, that is this FGF receptor 4 inhibitor, achieve objective response in 16% of the cases, as opposed to 0% in patients with FGF19 negative, and therefore this is a proof of concept. What do we know about precision oncology in cholangiocarcinoma? Well, these are the guidelines for management of cholangiocarcinoma. Our patient, again, will fall here, and Dr. Akulik also will expand on the potential treatment for this patient. But actually, in advanced stages of the disease, the conventional treatment is the combination of gencytabine cisplatin. So what type of information are we obtaining from a biopsy in this type of patient? Well, first of all, we will know the landscape of mutations. And I want to highlight that the genetic aberrations of intraparticulangiocarcinoma are completely different than the mutations of perihelial or distal cholangiocarcinoma, as you can see here. And these mutations may drive some therapeutic strategy. First of all, which are the most prevalent mutations in intraparticulangiocarcinoma? The most prevalent mutation is IDH on one hand, around 15% of the patients, and FGF receptor 2 fusions, 22% of the patients. And these two mutations are actually druggable. In fact, we have phase 2 studies. Here you have one in which objective response with FGF receptor 2 inhibitors achieved 20%, and this is another external confirmation of the same concept. Even it has been identified that FGF receptor 2 point mutations may be responsible of resistance to these agents. So in conclusion, regarding FGF receptor 2 are prevalent in cholangiocarcinoma, 25% of the cases. Objective response is achieved in 20%. Mutations in this driver lead to resistance to these inhibitors, and currently there are phase 3 trials comparing FGF receptor 2 inhibitors versus gemcitabine c-splatin. Finally, what about IDH mutations account for 15% of the patients, and this year there was this study, a single arm study, suggesting some activity when blocking IDH1. In reality, at ESMO, one month ago, was presented the first phase 3 positive trial with any drug in cholangiocarcinoma in terms of precision oncology. The trial was a phase 3 trial with a drug called ibosidinib that is an IDH1 inhibitor versus placebo in 185 patients in second line intraparticolangiocarcinoma that had IDH mutations, and there were significant differences in terms of progression-free survival with a hazard rate of 0.37. So if the patient has intraparticolangiocarcinoma at advanced stages of the disease, the benefit of having a biopsy in terms of precision oncology is that we might consider exploring FGF receptor 2 fusions or IDH mutations to receive these inhibitors. Finally, just a couple of minutes to talk about what liquid biopsies can do for a patient in the sense of HCC assessment and applicability. So first of all, liquid biopsy is a molecular analysis of tumor byproducts in the bloodstream. These products are circulating DNA on one hand, circulating cells on the other hand, and finally you have exosomes that may contain messenger RNA, miRNAs or non-coding RNAs. Of course, liquid biopsy is not invasive, is able to capture all cancer niches, the mutations, and is a dynamic exploration. The limitations are that the still is not completely optimized and not all tumor sites release sufficient quantity of DNA. Here I'm showing in one slide the opportunities for liquid biopsy in cancer. Cancer detection for screening and surveillance, molecular profiling for prognosis, detection of residual disease, monitoring response, and monitoring clonal evolution. In HCC there are very few studies. This is a study of our group, along with a group of Dr. Augusto Villanueva, in which we screen the tissue of eight patients. We identified 25 drivers, and in this pilot of study by liquid biopsy, we were able to recognize 71% of the biopsies, and some of those were drogable targets like JAK1 or VRAF. How I see the impact of liquid biopsy in HCC, in the evolution of the patients from cirrhosis, dysplastic nodules, early and advanced HCC, is that on one hand I think that this approach is very promising for early detection of HCC, and on the other hand is very promising for precision oncology, defining treatments, and also defining mechanisms of resistance. So in conclusion, the takeaway messages are as follows. Tumor biopsy is recommended in patients with cirrhosis and atypical liver mass. Patients with delivered M present in most cases intraparticular angiocarcinoma, or less so, atypical HCC. Tumor biopsy provides information on diagnosis, molecular features and prognosis, and precision medicine in advanced cases. Also not approved drugs yet are there. In hepatocellular carcinoma, we have FGF receptor 4 inhibitors in FGF19 positive patients, and in cholangiocarcinoma, we have FGF receptor 2 inhibitors for fusions of FGF receptor 2, or IDH inhibitors in IDH mutant cases. Biomarkers to predict response or resistance to checkpoint inhibitors in HCC is a clear unmet need at this point in HCC management. And finally, liquid biopsy has potential in HCC research for early detection and for defining precision treatments and resistance. Thank you very much for your attention.

precision medicine

liver cancer

tumor biopsies

molecular classification

systemic therapies

liquid biopsies