So we're going to move on to hepatocellular carcinoma now, and I'd like to invite Dr. Nihar Parikh to the stage. He's an Associate Professor of Transplant Hepatology and Medical Director of the Liver Tumor Program and Director of Hepatology at the University of Michigan. So over to you. Thank you very much, and I want to thank the organizers for allowing me to present today. My voice is a little off, so I apologize for that, but it's a cold. It's not what you think it is. So again, I'm from Nihar Parikh. I'm from the University of Michigan. I'm pleased to really present the HCC debrief. It's been exciting advancements in the field. So I've categorized this in different ways here, and you can see the abstracts that are included, including risk prediction, surveillance, prognosis, and HCC treatment, and we'll kind of go into depth on all of these different abstracts. So to start, we'll start with risk prediction for HCC, and the first abstract is from Vujkovic et al from the University of Pennsylvania. I just want to remind everyone that risk prediction and risk stratification has several benefits, and it's a hot field right now in HCC in terms of surveillance, because you can provide high-intensity surveillance for higher-risk patients, and there are additional benefits to minimize surveillance-related harms. You can do things like risk modification in high-risk groups, and these patients are prime for chemoprevention if you can identify them accurately. So in the first abstract, this is a novel genetic architecture for all-cause cirrhosis and hepatocellular carcinoma that was presented at one of the plenary sessions. The aims of this study were to really understand some of the, you know, augment the current repertoire of genetic variants that contribute to cirrhosis development and HCC development. They wanted to evaluate the interplay between these genetic loci and underlying disease ideologies, which we won't have time to talk about during this talk, and a really exciting part of this is to talk about genetic risk scores and how they could be predictive of outcomes in these cohort populations. For this, they used a very large, you know, set of data sources from diverse backgrounds of patients, and you can see here that they found eight loci that are shared between HCC and cirrhosis. Some of these were novel. Some of these were known, but there were also four that were very specific to HCC as well, and in the genetic risk score, when they developed this, they could really risk-stratify the population for development of HCC. That's what the axis here is on the Y-axis and the X-axis is over time, so you can see if you add these to the underlying liver etiology, you can increase your prediction of development of HCC. So they found 11 novel variants for cirrhosis and four for HCC. Interestingly, a lot of these were related to major histocompatibility complex class II alleles, suggesting some immune-mediated effects that may predict the risk of development of these, and they found this genetic risk score predicted the progression from cirrhosis, from chronic liver disease to cirrhosis and cirrhosis to HCC, and they found additional interactions with metabolic dysfunction with viral hepatitis and HCC risk as well. The next abstract is from Agrawal et al. This is really looking at patients with non-cirrhotic HCC, which is a significant population, but really understudied out there. This study shows that about a third of patients in the VA diagnosed with mastoid-related HCC had non-cirrhotic mastoid as an etiology, and so there's really sparse data characterizing the histological variants of non-cirrhotic HCC, so the group here wanted to really develop a kind of multimodal way of characterizing this using morphology-based molecular classifications, immunohistochemistry panels, and looking at somatic gene mutation profiles. So again, this is a rare population that's not been studied very well. This was from a group in India that included 40 patients for this analysis, the majority of which had mastoid as the underlying etiology, and again, they performed this multimodal evaluation. These are the top-line results, and I think what the take-home is here is that there's a lot of heterogeneity, but there can be some in the histotypes that we see, but there can be some, you can trace it back to some of the etiology and potential pathways that lead to HCC in these populations, and if you look at the genetics, I think the big message here is that TERT polymorphisms are a big driver of HCC, and it's true in this non-cirrhotic population as well, and so the conclusions here, again, where there is high heterogeneity, the TERT germline mutation is important. Microtubacular are predominant in mastoid etiology and this beta-catenin pathway signaling and male predominance suggesting a pathway for these patients to get this, and then this compact histotype, again, is a well-differentiated mutation, beta-catenin signaling, and about more prominent in females, so these are different histotypes. I think it's a small study, but important. This is a really difficult subtype of HCC to study. The next study will really talk about risk prediction and HCC in patients who've been cured of hepatitis C from Kanzaki, so there's a serum, it's called a serum biomarker score to evaluate HCC risk in HCV-cured patients with cirrhosis, so we do know that patients who've been cured of hepatitis C have a persistent risk of HCC. This was nicely shown by George Iwano at University of Washington in the study from the VA, showing that there's a persistent risk even after SVR in patients with cirrhosis, and so what the study team here was really trying to do is trying to develop a risk stratification pathway for these patients, and so what they did was they looked at tissue, derived a signature, and then they developed that. Based upon that, they developed a serum-based signature for HCC risk, and then they validated it. I'm going to show some of the validation data. This is the tissue signature validation, and you can see here that they've developed a really nice curve separation, and actually no HCCs developed in the lowest group. Again, small validation cohort, but powerful validation from the tissue signature. Now, when they looked at this in the serum, when they developed a secretome-based signature, they were, again, able to differentiate not quite as well as you would expect from the tissue, but you can see here that the factors that are involved, including some of the demographic factors of the patients. So this hepatic transcriptome-based PLS-HCV cure was developed and validated both in tissue and serum, and again, this reflects the molecular dysregulations associated with necroptosis and TNF-alpha activation. The serum proteome-based PLS-HCV cure was developed and validated. To what extent this can be used and implemented in clinical practice remains to be seen, but interesting data to see how we can risk stratify this high-risk group. The last risk prediction abstract is from the Binojohn at the Miami VA, which looked at liver stiffness as an integration into risk stratification in mass hold, and so they really wanted to look at the patients who've gotten liver stiffness measurements within the VA to establish cutoffs or trying to figure out subgroups within the MASL-D population that we can potentially identify as needing surveillance. So what they did was they used a large VA cohort, over 30,000 patients that had transient elastography and MASL-D, and they were able to identify some subgroups that meet the threshold for potentially being cost-effectiveness for surveillance. Again, males, LSM between 10 and 14.9, excuse me, and those with diabetes. So every five kilopascal increase in LSM increased HCC risk by 19%, and again, HCC surveillance may be cost-effective in males and diabetes, and so we can maybe integrate this in a way to personalize HCC surveillance in MASL-D. This is, again, a VA population. It'd be nice to see a more mixed population, women, et cetera, but I think this is a really nice proof of concept. I think the next topic we'll cover is surveillance for HCC. The first abstract is from Dr. Kim at University of Washington, and talking about the optimal AFP threshold for HCC screening should be lower than 20 for cured hep C, ALD, and MASL-D. We do know that serum AFP is a part of HCC screening recommendations currently integrated with ultrasound. Per the guidelines, AFP greater than 20 is considered a positive screening test, but we do know that AFP varies by liver disease etiology. In this study, they looked at 28,000 veterans, another large VA study, 1,000 cases, 27,000 controls, and they evaluated test sensitivity across etiologies of disease, and you can see here that activate hep C has the highest sensitivity for AFP, but at different thresholds, you actually don't lose that much in other etiologies of disease if you drop the threshold further from 20 to 10, and you may increase sensitivity a bit. Specificity is the other concern when you think about shifting your cutoffs, and actually specificity is relatively well-preserved in these patients who have cured hep C, ALD, and MASL-D. So AFP performs better as an HCC screening test in active HCV, which we knew, but in patients with cured hep C, ALD, and MASL-D, lowering that threshold from 20 to 10 would increase test sensitivity by about 10%, while maintaining that high specificity that you would want with a screening test. So I think this is a really nice study to show this, and I think a really large-scale way of evaluating AFP. The next study is about a novel way of testing or doing surveillance and using this multi-omics biomarker signature derived from circulating hepatocyte extracellular vesicles. Extracellular vesicles are an interesting way of performing surveillance. They're, again, circulating, and they are excreted by cancers, and they've been utilized in HCC and other cancers, but this is one of the first commercial companies to really come up with a way to look at this in a study. So these EV biomarkers, they have protein biomarkers off the EVs, mRNA biomarkers off the EVs, and then some soluble proteins, including AFP and DCP, and they were able to categorize HCC positive and HCC negative based upon this. They presented nicely some of the derivation work that led to the study, but in this abstract, they really focused on the testing set, which is in cohort two here, which included 16 HCC cases and 17 patients with cirrhosis, and you can see here the AUC was very nice, 0.9, very small case control, kind of a phase one study, but a 0.9 is great, and it compared favorably to GAD, which is DCP, AFP, gender, and age in this small study. If you look across stages of HCC, it performed very well in early stage disease with high sensitivity and specificity, which is what you want to see in these early studies. This approach needs to be validated in larger phase two and phase three studies, which I know the company is planning, but this is a really interesting approach for surveillance. Next one, we'll talk about prognosis from David Goldberg, who's in a category of his own, as many of you know here, talking about prognosis of patients with HCC. You know, this group was interested in developing an objective way to predict prognosis in patients with intermediate and early stage HCC to really think about how you can talk about overall survival, survival benefit of treatments, including liver transplantation, and existing ways that we do this right now are based on subjective criteria, including ECOG status, so they wanted to get rid of those subjective criteria. This is, again, another large VA study, retrospective cirrhosis cohort from the vocal group in patients diagnosed with HCC. They censored the patients at the time of transplant, but one of the, you know, good or bad things about the VA is that there wasn't a lot of censoring because there isn't a lot of liver transplantation. You can see the variables that were included here in the study, and these are the top line results which show that the AUC compared favorably to BCLC and ALBI score alone. They did combine BCLC and ALBI together, and it was very similar. Interestingly, 30% of the patients in this cohort were untreated, which, you know, I think they're going to, group is going to look at a little bit better, or a little bit more, but again, this was more predictive in that group specifically, which is relatively understudied. Again, this presents an objective score to predict survival from diagnosis in a diverse cohort with early and intermediate stage HCC. Good performance, not outstanding, but pretty good, and I think it's difficult to predict mortality in this group of patients, but again, these are potential applications. Thinking about expected survival, guided decision-making, and then combine these with post-transplant survival models to really think about what the survival benefit of transplantation is. The last two studies we'll talk about are HCC treatment. The first is from Ms. Bansal, who's a medical student and a second-generation ASLD-er here. So this is really looking at patients with child puberty, liver disease, and receiving first-line immunotherapy, which is the STRIDE regimen, Dervalumab plus Tremolumumab. So Tremolumumab in this regimen is given out at the outset as a CTLA-4 inhibitor. The STRIDE regimen is an approved first-line option in patients with unrespectable HCC. However, all the data in the registration trial were patients with child puberty. Child puberty patients, we see a lot in clinical practice. It's a relatively understudied population of patients, and so the group here really wanted to describe the utilization and outcomes in STRIDE in unresectable HCC and child puberty cirrhosis. They included patients from the VA. It was a small number of patients, but 30 patients is, you know, this is the first data in an IO plus IO therapeutic combination in child puberty patients that have been, you know, presented in a multicenter fashion. You can see most of the patients had a B7 disease, some with B8, and a small number with B9, and the majority had non-viral etiology of their HCC, which is also important because there has been some controversy about the utility of immunotherapeutics in non-viral etiologies of hepatocellular carcinoma. We look at the survival. They compared it to the patients who had CHOP-UA that received this. The CHOP-UA patients, the median survival was not achieved by the time of this, but we do see that the median OS was 6.2 months in the CHOP-UB patients. So not great, but certainly there are some patients that respond and have a prolonged survival. The other important part is how well was this tolerated? Grade two to three treatment-related adverse events happened in about 32% of patients. So actually pretty well when you look at the clinical trial data. And when they compared it to CHOP-UA patients, the rate of immune-related adverse events was very similar. So is this safe to give? Yes. Is it efficacious? I think it remains to be seen. You can see the numbers are pretty small as we get further out here. Greater than 25% of patients in the VA received STRIDE that were CHOP-UB, and two-thirds received prior LRT. Most of these patients had microvascular invasion or macrovascular invasion, excuse me, and extra hepatic metastases. Again, 76% were BCLCC. They got about two and a half cycles through, was the median time. And again, the median OS was 6.2 months. I did not show this data, but there was some deterioration in liver function, but it did not happen until later. And so some of this is probably just due to progression of their liver disease or their cancer, rather than a direct effect of the treatment. The last study was presented yesterday as a late breaker, as an encore from Josep Jovet from Mount Sinai. This is a LEAP012 phase three study of Linvatinib plus Pembrolizumab plus TACE for intermediate stage HTC. I think there's a lot of interest in combining therapies in earlier stage disease. We heard from GI ASCO last year of the Emerald 1 results with the primary endpoint being PFS. This is another combination therapy, therapeutic regimen to look at Linvatinib plus Pembrolizumab plus TACE in the treatment of intermediate stage HTC. It's a global study, randomized patients that were TACE eligible, one-to-one between Linvatinib plus Pembrolizumab plus TACE versus placebo plus TACE. There were various stratification factors that you can see listed here. And the primary endpoints were PFS and OS. And again, this is the first cut of the analysis. And so some of these are immature endpoints at this point. There are 480 patients assigned randomly. And you can see the flow charts here from the consort. The median time for randomization at this time of analysis was 25.6 months. So reasonable follow-up in this population. If you look at the primary endpoint of PFS, the study met its primary endpoint so far with a hazard ratio of 0.66. And you can see the confidence interval there. The median PFS in the patients who received the combo was 14.6 months versus 10 months who received placebo plus TACE. You can look at here, the objective response rates were also about 13% higher in patients who received the combination. Interestingly, the disease control, or excuse me, the duration of response were very similar here. So once you did get a response, the duration response was similar, but the absolute number of patients who had a response was higher in the combination. The disease control rate, again, you can see that is higher here. AEs are always concerned when we start adding therapies on here and the treatment-related AEs were a significant issue here, I think in the combination, including those grade fives, which we never like to see when we have patients on multiple therapies. And so I think as the data matures, we'll have to understand these patients this more. The serious AE event, again, was two to three times higher in the patients who received the combination. So in conclusion, this LEAP0012 study showed a clinically meaningful and statistically significant improvement in the PFS. I think the side effects were manageable. However, I do have concerns about the serious and the grade five side effects. So I think we'll have to see how this evolves. Not certainly ready for prime time in clinical practice yet, but I think we're gonna see in more and more studies like this in this space. So with that, I thank everybody. I thank the organizers for having us present and all the real presenters out there for sharing their slides with me to facilitate this. So thank you all. Thank you.

hepatocellular carcinoma

risk prediction

genetic architectures

prognostic models

treatment strategies