™ presentation titled, “Evolving Approaches for HCC in the Molecular Medicine Era.”

We are witnessing a change in the treatment paradigm for advanced hepatocellular carcinoma [HCC]. The emergence of immune checkpoint inhibitors in this space has opened the door to many exciting new questions on how to treat this difficult disease. Today, we are going to discuss how modern molecular approaches are helping to expand the therapeutic options for patients with liver cancer.

I’m Dr Richard Finn, from the David Geffen School of Medicine at UCLA [the University of California, Los Angeles]. Joining me today are my colleagues: Dr Catherine Frenette, from Scripps Health in La Jolla, California; and Dr Michael Morse, from Duke Cancer Institute, Durham, North Carolina.

Thank you so much for joining us today. Let’s dive right in.

The past few years, as mentioned, have been a really exciting time for those of us in liver cancer research, and also for our patients who have been in need of new, more effective treatment. Liver cancer is a complicated disease. I say that not to be self-serving, as a liver cancer researcher, but I say that because it’s really 2 diseases in 1, right? It is a malignancy and the complications of having tumor burden, and then it’s also almost always accompanied by underlying liver disease, chronic liver disease, cirrhosis. That is why a multidisciplinary approach is really of benefit.

As you see on our panel today, both Michael and I are medical oncologists, and Catherine is a hepatologist with interest in liver cancer. I want to open up the discussion with you, Catherine. Give us some thoughts on how you approach patients with liver cancer in the context of underlying liver disease, and give us insight into that multidisciplinary approach.

 Thanks, Rich. I’m really honored to be on this panel with 2 great oncologists who are very esteemed. Hepatology has been involved in liver cancer for years. As you said, a lot the patients have underlying cirrhosis. So as a hepatologist, I feel like my job is to keep the liver as healthy as I can so the patient can continue to receive therapy for their cancer. Whether that’s locoregional therapy, surgical therapy, systemic therapy, or whatever therapy they’re going to get, I need to keep that liver as healthy as I can. That might involve discovering what the underlying cause is and may involve treating that. For instance, getting them on hepatitis C treatment or hepatitis B treatment. In patients with cancer, it’s been shown that treating the virus can actually improve their outcome. Or, getting their encephalopathy, or their ascites under control and keeping that under control while they’re undergoing their systemic treatment.

It really requires a multidisciplinary evaluation and care. We bring in hepatology, oncology, surgery, radiology—all of these different specialists—to provide the best care to patients.

Optimizing Care for Patients With Hepatocellular Carcinoma

 We talk about a multidisciplinary approach for the disease. The reason we do that and involve all of these various players is that depending on the stage, there are different modalities used. When you approach a patient, Catherine, how do you stage them?

 Staging for liver cancer is very different than it is for other cancers, as you well know. I always teach my Fellows that when we’re talking about colon cancer, gastric cancer, or we’re talking about the TNM [tumor, nodes, metastases] staging system, it’s all about how much tumor is there and where it is. In liver cancer, we have to take into account the underlying cirrhosis, the performance status of the patient, and then we talk about how much cancer is there and where it is.

I really like the BCLC [Barcelona Clinic Liver Cancer] staging system because it really does go through that mental exercise of, how bad is their liver disease? What’s their Child-Pugh classification? What’s their performance status? Then, how big are the tumors? Where are they? Do they have metastatic disease, vascular invasion? This leads you down the path of various treatments. That’s really important when we’re thinking about staging, and the best staging systems for liver cancer, all of them include liver function as part of the staging process.

 In medical oncology, we stage patients for several reasons. One reason is so we’re all talking the same language, right? When we look at data from a study, or we’re talking to our colleagues and we say a specific stage, we know what is being talked about and can compare studies. And it’s also prognostic. We can tell a patient, or at least have some assessment in our own mind of how they will do.

In the context of the BCLC staging system, we know there’s very early stage hepatocellular carcinoma. Generally, these are patients with small tumors, preserved liver function, who could potentially be cured with dissection or radiofrequency ablation. Or, patients who might have early stage disease but poor liver function, who might be cured with transplant.

And then there are the patients who have very advanced, decompensated liver disease. For these patients, hospice and supportive care might be best; or even a liver transplant if their tumor burden is acceptable.

And then we have the big groups of patients with intermediate disease and advanced disease.

Michael, Catherine and I are both on the West Coast, you’re on the East Case—at Duke Cancer Institute. At Duke, in your multidisciplinary program, do you tend to use BCLC? Also, for the audience, could you distinguish between intermediate disease versus Barcelona stage B, versus Barcelona stage C?

 Yes. This is really where the liveliest discussions are likely to occur. As we get better and better therapies, we are seeing the suggestion that we could, perhaps, be treating some BCLC stage B patients, the multifocal disease, with systemic therapies. At the same time, I know our interventionalists certainly feel that there are some BCLC stage C patients that they can approach with something like Y-90 [yttrium-90] radioembolization.

We obviously have had some studies that have tried to compare these modalities—the SIRveNIB and SARAH trials. You interpret them based on what you know about managing the disease, where you come from, what tools you have. I would say that the really exciting direction, this isn’t standard of care yet, but it’s certainly where the research is going, is probably the need to combine these approaches.

Right now, you have to have the discussion regarding the standard of care. Is this person going to be a candidate for locoregional therapy, or are they going to get systemic therapy? In the future, we might talk about combining them.

Hepatocellular Carcinoma Risk Stratification and Therapy

 In liver cancer, we now stratify based on clinical characteristics—tumor burden and liver function. And obviously, patients will have their greatest survival, as Catherine alluded to, if they have well-preserved liver function. And so, it’s important that we triage patients appropriately to the correct treatment at the right time.

Someone might start with intermediate disease, liver confined disease, no vascular invasion. They may get locoregional treatment, and then they might progress and become advanced. I always emphasize that patients can have advanced liver cancer and still have tumors confined just in the liver, right? They might have a tumor that’s invading the vasculature, or what we call macrovascular invasion. Or, they might get locoregional treatment, as you alluded to, and recur very quickly. Or, they receive repeated locoregional treatments to their tumor, and every time they get another scan there’s still residual disease, and they have some resistance. Those would be patients who would be candidates for medical treatment.

As we have more medical treatments, systemic treatments, I think we need to reconsider the limits of locoregional therapy. But still, when we talk about cancer medicine today, we talk a lot about molecular profiling. We talk about cancer genomics. We talk about PD-1/PD-L1.

Michael, I know you’ve always had some laboratory interests as well as clinical interests. Has this personalized approach impacted liver cancer? What do you see the treatment landscape including when you see a patient? Is it still a one-size-fits all approach, so to speak?

 Well, outside of AFP, we don’t really have a biomarker that we use. We use immunotherapies regardless of PD-L1 status, for example.

Obviously, there are molecular differences in HCC [hepatocellular carcinoma] cases. There’s recent work that’s been done with circulating tumor DNA that demonstrates that there are at least 2 separate groups—there are probably more—that have certain molecular aberrations. Most of those molecular aberrations are not truly actionable, unfortunately, and in regard to those few that do occur that are considered actionable in other malignancies, we don’t really know how to target them in hepatocellular carcinoma.

I do think it’s a start. We have the tools to be able to look at molecular changes, whether it’s by circulating tumor DNA or biopsying the tumor. I think as we do this more and more in clinical practice, we’ll find some subgroups that might be treated differently. But right now, the simple answer is that we treat everybody fairly similarly. AFP is the 1 marker we have for 1 particular drug.

 Yes. Some of you have said, “Well, you don’t biopsy people with liver cancer.” Maybe that set us back a little bit? Although there are some alterations that occur with various frequency, many of them are not actionable. There has been some interest in the 

 amplification and targeting fibroblast growth factor receptor 4, which is its known target for 

The landscape and the treatment of liver cancer is changing so fast. I think we haven’t seen where some of these niche indications will go. For example, c-Met was looked at a while ago, and there are some mixed signals about c-Met expression versus amplification being a predictive marker for response to targeted therapies against the HGF axis.

Catherine, do you biopsy all of your patients with advanced liver cancer?

 It’s interesting how the field has moved. And you’re right, Rich. I think it might have set us back a little bit. Years ago, we used to biopsy patients. Then, we really got away from doing so because we didn’t have markers that were useful. You’d see a few complications and get a little bit anxious about biopsying people when you’re not going to do anything different with the information, necessarily.

I think that now with where the field is moving, we are going to. We are getting to a point where we may have some useful analyses.

I have started biopsying more patients. I definitely agree, now with the move of all of the studies requiring biopsies again, that it’s important to look for these targetable mutations and try to identify how to group patients into categories for treatments that may actually work for them.

 If I could take a slightly different tack? We’re actually really excited about circulating tumor DNA analyses at our institutions. We’re building them into many studies now.

 looked at more than 200 patients. Almost 90% of the patients had some molecular aberration detectable. They weren’t always actionable, but I think this says that hepatocellular carcinomas typically do release DNA into the circulation. And as I think we get more comfortable with circulating tumor DNA in other cancers—lung cancer, for example, would be a prime example—I think that could carry over into HCC, where we’ve been so reluctant to biopsy for fear of causing complications.

And once we accept circulating tumor DNA as giving results that are as good, if not better than the biopsy, I think we can really argue that most patients should be given the chance. Every once in a while you do find an amplification. We don’t know exactly what that means in HCC, but it allows patients to be put into basket trials so we can learn whether they’ll respond to these targeted therapies.

 occurs in a single-digit number of cases, but there are some instances. I’m sorry, Catherine, you had a comment?

 I was going to say that I agree that circulating tumor DNA is incredibly interesting. It may end up being more accurate in HCC than a biopsy. These tumors can be so heterogeneous. I’ve seen 1 tumor have 3 different types of tumor within it, all could be releasing their circulating tumor DNA. That could give us a lot more information than a single needle biopsy, potentially. I do think this is exciting.

 Yes. I think the true impact will come, hopefully, in the near future.

Testing for Targetable Mutations in Hepatocellular Carcinoma

 When we talk about validated targets in liver cancer, I think the one that sticks out in my mind is VEGF. You all can certainly weigh in on that. That is my personal opinion.

Before we had sorafenib, there were data evolving about the importance of vascular endothelial growth factor [VEGF] and the VEGF receptor in liver cancer. On imaging, liver cancers tend to be hypervascular. One of the treatments here is chemoembolization, right? That embolization component where we block off the blood supply to the tumor certainly can help people live longer and control these tumors.

And then the first drug to be approved in liver cancer was sorafenib. Sorafenib, like other drugs, is somewhat of a dirty kinase inhibitor. It clearly targets the VEGF receptor. It was approved in kidney cancer as a proof of concept. Then it was looked at in liver cancer versus placebo in the SHARP study, as well as a companion study done in the Asian-Pacific region. Both established sorafenib as the standard of care for advanced liver cancer, or liver cancer that had progressed.

That was based on data from 2007 and 2008 published in the 

 that showed that sorafenib improved survival versus best supportive care on a median of about 3 months. That translated into a hazard ratio of about 0.69. And interestingly, it did that without inducing response. It improved survival by slowing progression, and PFS [progression-free survival] was significantly improved. It didn’t necessarily improve quality of life, but it didn’t decrease quality of life despite the fact that it has well known [adverse] effects in this population—the hand-foot syndrome; GI [gastrointestinal] toxicity, specifically diarrhea, anorexia, weight loss. There is some hypertension, but I don’t think that has been a big challenge when treating these patients.

I remember being interviewed in 2008. “Sorafenib is here. What do you think?” We thought, “Well, this is the golden age of liver cancer. We’ve proven that systemic treatment works. There will be a number of new studies and new drugs approved.” I was right, but I was off by about a decade. We went through a decade where we had no new drugs.

And then along came lenvatinib. Before I talk about the clinical data with lenvatinib, Michael, again, you and I were involved in a review article in 

 about a year ago that explored the targeting of VEGF. Can you provide us some of your insight in regard to the strengths and weaknesses of targeting VEGF?

 Well, as you pointed out, these are vascular tumors. It has now been proven with multiple drugs that have VEGF receptor TKI [tyrosine kinase inhibitor] activity that there’s activity in targeting that pathway. However, you’re also right that the drugs that have been most successful are those that are dirty kinase inhibitors. Which of the other targets is also relevant, or whether any of them are truly relevant, we don’t really know.

But the other issue is it’s also clear that the strength of the inhibition may affect activity. Lenvatinib inhibits VEGFR2 more effectively, or at lower concentrations, than sorafenib would. So to me, that also proves that it’s important to target VEGFR2, and then, perhaps, other pathways as well.

Targeting the VEGF Pathway in Hepatocellular Carcinoma

So lenvatinib comes along. It’s a very potent VEGF receptor inhibitor, but also has pretty broad activity against the fibroblast growth factor receptors—FGFRs 1 through 4. We talked about specific alterations in the FGF pathway, and molecular alterations that occur in liver cancer. But, there was a lot of interest for a while in targeting FGF as a resistance mechanism to VEGF, right? There was a drug, brivanib, being developed in the past that was felt to be a very attractive agent because it not only hit VEGF but it hit FGFRs as well.

The studies with brivanib did not meet their end points. However, we did see a head-to-head study, REFLECT, which was a pivotal study that got lenvatinib approved. It was a global study of over 900 patients. This was an open-label trial of sorafenib versus lenvatinib in the frontline setting. Again, as with all of our clinical studies, these were patients who had Child-Pugh A cirrhosis, well compensated. This head-to-head study was powered for noninferiority. We saw that lenvatinib actually met its end point. It did not meet the superiority end point. 

Survival was around 13 months with lenvatinib, and about 12 months with sorafenib, and that translated into a hazard ratio of 0.92, I think. But the confidence interval overlapped 1. So this was not statistically significant for superiority, but the upper limit of that confidence interval was 1.06, which met the noninferiority end point—being under 1.08.

Based on these data, lenvatinib did get approved. What was interesting in the data set was that lenvatinib met its secondary end points. It improved progression-free survival versus sorafenib. Essentially, it almost doubled it. And also, we saw a reasonable response rate with a TKI [tyrosine kinase inhibitor. 

The objective independent review RECIST response rate was just about 19% with lenvatinib. Historically, we saw sorafenib had some single-digit responses. And if you look at the modified RECIST response rate, which takes into account some of the enhancement characteristics, that response rate is around 24%. 

We also know that lenvatinib and sorafenib have different adverse effect profiles. Lenvatinib has more of a hypertensive effect, and also the VEGF component of more proteinuria. Both drugs have similar amounts of GI [gastrointestinal] toxicity, including diarrhea, anorexia, and weight loss. However, sorafenib has more hand-foot skin syndrome. We really don’t see a significant percentage of hand-foot skin syndrome with lenvatinib, which I think is a differentiator that we consider when approaching our patients. Also, keep in mind the REFLECT study excluded patients who had tumors occupying more than 50% of their liver, or invasion of the portal vein outside the liver. But needless to say, we have 2 TKI options in the frontline setting that have some subtle differences.

HCC: VEGF TKI Data and Treatment Selection

 Now, the big change in all of oncology, and I would say even more recently in liver cancer, has been the impact of PD-1 and PD-L1 agents. Most recently, we saw some of the data with atezolizumab and bevacizumab. That’s based on the science of targeting PD-1, PD-L1, and VEGF. Michael, again, leaning on some of your preclinical experience, could you describe the background for targeting these 2 pathways?

 The interesting observation across malignancies, as you pointed out, is there can be tumors that frequently have T-cell infiltration. Those T cells are poised to attack the PD-1, PD-L1 interaction. And whether that PD-L1 is deployed by the tumor cells or other infiltrating immune cells, it prevents the T cells from being able to kill. It causes the T cells to be dysfunctional. And that could be relieved by a blockade, either at the PD-L1 or the PD-1 portion of the interaction.

What’s even more interesting, though, is that if you look at anti-VEGF strategies—whether it’s hitting the receptor, whether it’s finding VEGF—it seems to modulate the immune system as well. We actually did some work, as did others a number of years ago, looking at dendritic cells. These are the key stimulators of T cells in the body. We looked at whether their maturity level was changed by inhibiting some aspect of the VEGF pathway. Indeed, they do seem to be suppressed when there are high levels of VEGF. They mature. They’re more effective stimulators of T cells when you inhibit VEGF signaling. And so, there’s a rationale for combining an anti-VEGF strategy with immunotherapy. Obviously, that’s been borne out now in several malignancies.

PD-1/PD-L1 Antibodies and VEGF Inhibitors in HCC

 In liver cancer, we recently saw the approval of atezolizumab, a PD-L1 antibody, in combination with bevacizumab, an anti-VEGF antibody, based on the IMbrave150 study that was published in the 

 in May of 2020. The drug was approved by the FDA a few weeks later.

For the first time, we have a regimen that improved overall survival [OS] versus sorafenib, right? We had commented that lenvatinib and sorafenib were noninferior. However, we actually saw very significant improvement in overall survival, with a hazard ratio for OS of 0.58; improvement in PFS [progression-free survival], with a hazard ratio of 0.59; and response rates of about 27%—objective response rates.

I have lived through this era of metastatic drugs, with multikinase inhibitors and even the single-agent PD-1 inhibitors—nivolumab and pembrolizumab, which have approval in the second-line setting based on accelerated approval mechanisms—where we see response rates of around 15%, or 18%, 20% at best in the second-line setting. But now, we have a regimen that’s giving objective responses over 27% and actually shows improvements in quality of life. But like always, the regimen has some safety concerns.

Catherine, you’re a liver doctor, and we’re talking about antibodies. One of them can cause autoimmune phenomena, and one can cause bleeding. How do you look at the atezolizumab/bevacizumab data as a hepatologist, and regarding safety concerns?

 I was actually very impressed with the safety that I saw in the studies. They were very careful about choosing patients. Every patient had to have an endoscopy within 6 months, and any varices that were high risk had to be treated per local standards. That was very important, I think, to maintain safety and show that bevacizumab, a VEGF inhibitor, can be used safely in patients with underlying chronic liver disease. We know it is associated with increased bleeding risk, and we are aware of the wound healing issues, of course.

If you think about it, all of the TKIs [tyrosine kinase inhibitors] have VEGF as part of their actions. Every patient with cirrhosis and liver cancer needs to have an endoscopy for variceal screening anyway. In addition, all of the TKI studies actually needed an endoscopy as well. So this was not an unusual part of this study, but I think it’s a big discussion just because of the historical experience with bevacizumab that we’ve seen.

I do worry about patients with underlying liver disease developing immune-mediated adverse events, whether it’s autoimmune hepatitis, or even autoimmune nephritis or pneumonitis. A lot of times, these patients are a bit fragile. Developing an immune-mediated adverse event can sort of make them fall off the cliff, so to speak. So I do have anxiety about it. I do think they need clear education and need to be provided with warnings about starting these drugs. “Hey, you may develop this. It may result in decompensation of your liver. It could even potentially result in death. You may need steroids.”

However, the percentages that we saw were very low. In the atezolizumab/bevacizumab study, fewer than 10% of patients actually developed autoimmune hepatitis. And a very low percentage actually had to stop therapy because of immune-mediated adverse events. I think it was around 10%.

That was very promising. Obviously, when we talk more in this discussion about the combination of the PD-1, PD-L1, and the CTLA-4s, those combinations have quite a bit more toxicity in terms of immune-mediated events. We have to have a further discussion about those, with regard to safety. But in general, I’m very comfortable with the safety of atezolizumab/bevacizumab, at least in patients with Child-Pugh class A cirrhosis.

 Yes. When you lined it up to its active control, which was sorafenib, it actually had a very favorable safety profile.

Atezolizumab/Bevacizumab Data in Frontline HCC

 Bevacizumab and the risk of bleeding has always been a concern. What types of patients are considered high risk for bleeding? I don’t mean outside of the GI [gastrointestinal] tract. I specifically mean someone who has an endoscopy. Who’s at high risk of bleeding?

 Patients at high risk for bleeding would include those with large esophageal varices, or large gastric varices. Also, sometimes with esophageal varices, we grade them—grade 2 or grade 3. Those patients are at risk for bleeding. We also look at what we call high-risk stigmata, which are red wale signs. I explain to my Fellows and medical school students that a red wale sign occurs when the wall of the vein is so thin that you can see the blood through it. That’s a high-risk stigmata, and that’s a bit scary. Those are the patients we worry about.

We do know that patients who have more advanced liver disease—Child-Pugh class B and Child-Pugh class C cirrhosis—are at higher risk of bleeding, even with small varices. But generally, most people have not yet been putting those patients with more advanced liver disease on atezolizumab/bevacizumab because of that concern of bleeding.

 If I’m considering a patient for systemic treatment, at this point I need to come up with a reason why I wouldn’t give them atezolizumab/bevacizumab. The regimen is so active, and it’s improving overall survival. To me, it probably would be considered the frontline standard of care.

Michael, what are your thought about its place in our management?

 I would agree. In general, that’s true. Remember, as Catherine pointed out, these are Child-Pugh class A patients who were studied. We don’t know what the impact is going to be, frankly, in my practice the more common are the Child-Pugh class B patients. We also don’t know about the people who may start out as Child-Pugh class B or Child-Pugh class C but see Catherine and have their liver disease managed and slide back into a better Child-Pugh classification. We don’t know how those people are going to fare.

I think while this is relatively new and people are getting experience with it, it’s best to understand the Child-Pugh class A patients are those for whom activity was seen. We don’t know what the activity would be like in others and what the safety data are.

Obviously, there are some other subgroups of patients for whom we would all agree are not appropriate candidates, for example, those who need a liver transplant. At our center, we’re certainly not willing to give a checkpoint blockade in that scenario. There are a few patients who have other very severe autoimmune diseases for whom we have to be careful. There are some patients for whom we can’t control their varices.

Catherine, I’m interested in people with gastric varices, because those sound like they’re more difficult to manage than esophageal varices?

 They very much are. Esophageal varices are more of an up and down tube. When you band those, they scar down nicely. Generally, they do well. There’s always a risk of post-banding ulcer bleeding, which generally occurs 7 to 10 days after the banding event. So you have to wait and check.

The gastric varices are more of a bunch of grapes, in regard to configuration. When you band gastric varices, oftentimes when you get that post-banding ulcer it can erode into the next varix and actually worsen bleeding. It’s very scary, and they can be quite difficult.

We’re coming up with other ways to treat gastric varices. We’re doing EUS [endoscopic ultrasound] glue or coil injections. We’ve done something called a balloon-occluded retrograde transvenous obliteration, where we actually can sneak up through a splenorenal shunt and embolize them from the inside. Obviously, that requires a specialty center, but gastric varices can be very difficult.

Atezolizumab/Bevacizumab for HCC: Toxicity Management

 We have some consensus that we have a lot of options in the frontline setting. I think atezolizumab-bevacizumab stands out because of its significant survival advantage over TKIs [tyrosine kinase inhibitors]. Single-agent PD-1 agents have not succeeded in the frontline setting, although we have seen some safety data with nivolumab and Child-Pugh class B patients. Again, it’s uncontrolled data, but the response rate seems to be maintained and the safety seems to be acceptable.

There are also some interesting studies about antidrug antibodies, right? It’s being reported that the incidence of these are a little higher with atezolizumab than with nivolumab or pembrolizumab. At least from my standpoint, I don’t see the clinical relevance of this just yet. We have data that clearly show that atezolizumab-bevacizumab is effective and superior to sorafenib. We have no clinical tests for these antibodies. And whether that will be responsible for acquired resistance has yet to be seen.

If a patient is not a candidate for atezolizumab-bevacizumab for some of the reasons we’ve discussed, Catherine, how do you decide between lenvatinib and sorafenib?

 That’s a really good question, and that’s where experience comes in and patient decision-making and knowing your patient. I have a discussion with my patient about their other medical problems and their life. For instance, if someone has a history of cardiac disease and they’ve got severe hypertension and are on 3 antihypertensives, I’m a bit more reluctant to give them lenvatinib. If they do get hypertension, it’s going to be difficult to control it when they’re already maxing out on medical therapy.

On the other hand, if someone is an avid gardener or an avid golfer, and a hand-food-skin reaction would really affect their ability to do what they love to do, I’m more reluctant to put them on sorafenib. I look at those kinds of things. I also think about the proteinuria. For instance, some patients have a really hard time with the twice-daily medicines, in regard to remembering to take the medicine twice a day. In that case, I might lean more toward lenvatinib because it is only administered once a day. You have to pick out these really little particularities of each patient in making the decision.

 Sure. The toxicity comes into play. And then sometimes, also, the activity, for those patients who might need a response, right? We know there’s a good chance of response with I/O [immuno-oncology], but this patient, because of 1 of the reasons mentioned, should not get I/O or a combination, specifically atezolizumab-bevacizumab. But they need a response. They have a bulky tumor. Sometimes that would make me consider lenalidomide over sorafenib.

Choosing Between VEGF TKIs for Frontline HCC Treatment

 Michael, we have other TKIs [tyrosine kinase inhibitors] in liver cancer, specifically in the second-line setting—regorafenib and cabozantinib—that were approved based on randomized data vs placebo. Both had similar hazard ratios—in the 0.6, 0.7 range. They’re both VEGF receptor inhibitors, plus. Specifically, cabozantinib has activity against c-MET and AXL, which is unique among the TKIs. But again, their adverse-effect profiles are fairly similar.

And the approval of atezolizumab-bevacizumab in the frontline setting has changed or disrupted our phase 3 data sets, right? None of the second-line drugs was studied after prior treatment with atezolizumab-bevacizumab. Even the approval of lenvatinib disrupted that a bit. They all followed sorafenib.

We have ramucirumab, as you commented on earlier, for patients who have an elevated AFP. That drug targets VEGFR2. We also know that bevacizumab targets VEGF. So a patient gets atezolizumab-bevacizumab in the frontline setting. At some point, they have progression. Michael, what are you going to do?

 Thank you for picking on me. The simple answer is, we don’t know. We have to do something for those people. For somebody who has maintained their performance status and we are able to manage the toxicities of these drugs, it’s fair to say we should start them on something and, if they progress, move on to the next drug.

You could make the argument that cabozantinib was tested in patients as either second- or third-line therapy, and approximately 7% or so of the patients who were on the CELESTIAL trial had received a prior immunotherapy. I don’t think we have that data with regorafenib, mainly because of when it was studied.

It is also a purely post-sorafenib study.

 Correct. You could make those academic arguments about what we do when we don’t have data. Cabozantinib was studied in a slightly broader patient population, without the requirement that, for example, a lot of people had more liver involvement, with portal vein involvement. And you could have bile duct involvement. But the reality is, both drugs—regorafenib and cabozantinib—have activity. And so I offer both options to people.

The main point that I’m trying to make here is that it is important that we manage those toxicities so people can get the benefit of the drug. And if they progress, hopefully they’ll be a candidate for third-line therapy.

With atezolizumab-bevacizumab leapfrogging to the frontline setting, it sounds as though you go to 1 of the second-line drugs—regorafenib or cabozantinib—when patients progress. Do you see lenalidomide or sorafenib being used in that setting?

 That would be very reasonable. As you know, the NCCN [National Comprehensive Cancer Network] Guidelines do include sorafenib in that scenario. They are all reasonable options. Ideally, we’ll have some data in the future, although I’m not sure we ever will, looking at what that proper sequence should look like.

I suppose you could even extend that to changing the immunotherapy component or doubling up on the immunotherapy component. Those are legitimate questions, even where ramucirumab fits in a patient who’s already had bevacizumab.

Catherine, what is your take on this? It’s expert opinion, so…

 It’s tough, and you’re right. There are no data, and it’s really anybody’s best guess at this point. In general, we would lean toward the second-line options: cabozantinib, as Michael had said, or regorafenib, potentially. I probably would not go to ramucirumab, even in patients with an elevated AFP. That said, I’ve used lenvatinib and sorafenib in the second-line setting. It’s very muddy waters at this point, so it can be a little difficult to make that decision.

Yeah. I was part of a review that we published early this year. Many of us thought you would go to the frontline TKI because of the choices we talked about before. Then we would look at regorafenib or cabozantinib, which would become third-line therapy because they’ve been studied after a prior TKI. Needless to say, we have made a lot of progress in liver cancer and need to adapt that data into the settings of progress we’ve made. We shouldn’t throw out all this data. We’re never going to repeat phase 3 studies just because the frontline choice came forward.

What we have proven from the phase 3 studies is that the drugs we have available are active antiliver cancer drugs. It’s important that we give patients the option to get these. That really brings up some clinical management questions.

First, we can’t wait until a patient has clinical progression anymore, right? When all we had was sorafenib, imaging became less important. Now that we have other drugs available in the second- and third-line settings, following patients with scans and documenting radiographic progression before they decompensate or have clinical progression is important, to give patients the benefit of these other drugs that appear to be active.

That also relates to something we spoke about earlier: transitioning patients, right? From intermediate to advanced, from locoregional treatment to systemic treatment, this needs to be done in a careful, appropriate, and thoughtful way, so we maximize the opportunity to get systemic drugs. I’m starting to see patients who have large tumors, who have less compensated liver disease, who get a response and their livers actually get better. Their liver function improves because they responded. This is a whole new era for us in liver cancer research.

The other thing, Michael, is that now we have ipilimumab-nivolumab approved in the second line, right? We saw an accelerated approval of CTLA4 plus PD-1. Is there a precedent to use that after prior progression on a PD-1 or PD-L1 inhibitor?

 Certainly it’s a new concept in hepatocellular carcinoma [HCC]. We really don’t. If you look at some melanoma literature, there can be activity. It just points out the fact that we don’t really understand which patients are likely to benefit from the immunotherapy, and what the resistance mechanisms to the immunotherapy are in hepatocellular carcinoma. In a nonrandomized fashion, it does appear that there is more activity for anti–PD-1/anti–CTLA4 than single-agent anti–PD-1 therapy. Again, there were different cohorts of patients in CheckMate040, and we have data on durvalumab and tremelimumab as well.

But whether that warrants the use of those, as Catherine pointed out, at least with nivolumab and ipilimumab having toxicity issues—particularly with ipilimumab 3 mg/kg— we have to be very careful regarding which HCC patients we’re applying them.

HCC Treatment After Previous I/O Therapy

 You touched on a few things. As patients are better compensated, getting started on systemic treatment at the right time, they are going to cycle through many drugs, outside of going on a clinical trial, which is always the best option. In practice, patients will get more drugs based on what we have available. The options probably include I/O [immuno-oncology], or immune checkpoint inhibitor combinations after prior I/O, in reality.

But you’ve commented on some interesting things that are in development. Ipilimumab-nivolumab, which got accelerated approval in the second-line setting, is now in the frontline setting. This is based on the CheckMate 9DW study, which explored the option vs sorafenib. At the same time, we saw the concept of CTLA4 and PD-1, PD-L1 inhibition based on the HIMALAYA study. That’s durvalumab and tremelimumab vs sorafenib, and single-agent durvalumab. This was a multiarm study. And there was data at ASCO [American Society of Clinical Oncology Annual Meeting] this year from Dr Robin Kate Kelley that showed response rates of the high 20% range with that regimen, similar to what we saw with nivolumab-ipilimumab in the second-line setting—with response rates of around 35%. These I/O–I/O combinations look very promising, and we’re waiting for phase 3 data.

At the same time, building on this idea of VEGF inhibition and PD-1 inhibition, we have lenvatinib and pembrolizumab. We published a phase 1b study in the 

 in September of this year. This was a nearly 100-patient study that showed this combination had a response rate of about 36%, which is very similar to what was seen in the atezolizumab-bevacizumab single-arm phase 1b/2 study.

Many of these responses are very durable. The adverse-effect profile looks very similar to single-agent lenalidomide or single-agent pembrolizumab, with no new synergistic toxicities. The overall disease control rate, including stable disease, was around 88%. It was very high.

The LEAP-002 study has completed accrual. This is a placebo-controlled study of lenvatinib plus pembrolizumab, or lenvatinib alone. We’re waiting for the results of this study, just like we’re waiting for the results of HIMALAYA that also has completed accrual.

And then there’s the COSMIC-312 study, which looks at moving cabozantinib to the frontline setting. Here we see cabozantinib combined with atezolizumab vs cabozantinib alone. There haven’t been much data presented with this combination, but they’re building on the consistent theme of VEGF inhibition, or multikinase VEGF inhibition, in combination with immune checkpoint inhibition. This tells me that the landscape has a lot of potential to change in the coming year or so, depending on when these events read out. It’s a very exciting time.

With these studies ongoing, waiting for those readouts, Catherine, what do you see as an unmet need for our patients with advanced liver cancer?

 One of the places we really need to have the data is when we consider moving these amazing systemic therapies into an earlier-stage patient. One of the discussions we always have at our tumor boards, as Michael pointed out, is, “Well, we can take that,” “We can do an ablation,” or “We can do a Y90.” Okay, but that also is going to take out some healthy liver. We certainly don’t want to decompensate someone with locoregional therapy over and over. Maybe we need to be moving to systemic therapy after 1 locoregional therapy if they didn’t get the response we wanted to see. Maybe we need to add these systemic therapies as adjuvant therapy after an ablation or resection. It would be amazing if we could decrease the recurrence rate after a resection and cure more of those patients, which would then encourage more patients to go to resection as opposed to going to transplant. We now know that they have a better chance of being cured compared with resection. Maybe 2 or 3 lesions should be resected, because the recurrence rate can be dropped down with these combinations? We have some really exciting opportunities to really change the whole way HCC [hepatocellular carcinoma] is treated.

Novel Combinations Under Study for HCC

 Looking deeper into the pipeline, Michael, looking into your crystal ball, are there targets or mechanisms that are waiting to be exposed in the liver cancer space that you’re interested in?

 We talked earlier about the molecular makeup of these tumors and how there are some targetable and actionable mutations. We have to find them. We have to test for them. The activity of drugs against some of those targets will be found in basket trials. We can’t forget that there are some patients for whom we do have drugs that could be beneficial. It’s a small number, but they’re there.

The second area you pointed out is FGFR4, which is very different from FGFR1, FGFR2, and FGFR3. You alluded to the fact that there can be amplifications of portions of chromosome 1, I believe, where FGF19 sits.

 Chromosome 11, I’m sorry. Thank you. There’s a small percentage of those patients, and perhaps targeting FGFR4 will have activity. A higher number of hepatocellular carcinoma patients have elevated levels of FGF19 expression but don’t happen to have the amplification. So that’s kind of an example of some of those needles in the haystack that we haven’t really addressed.

Of course, I have to say, immunotherapy continues to be an area where there should be more expansion. There are T-cell products that are being looked at in hepatocellular carcinoma because they are in many other solid tumors. There are also bispecifics being tested in this disease.

 Also, cell-based therapies, right? There are technologies with mechanisms that target specific proteins in liver cancer. Adaptimmune Therapeutics plc is developing a cell-based therapy against AFP-expressing tumors, right? That obviously has a lot of promise. As we’ve seen in other tumor types, harnessing the immune system with specificity can lead to improved outcomes for patients. And also, whatever works in other cancers for overcoming resistance to immune checkpoint inhibitors will probably have an impact in liver cancer.

But when I look at lung cancer and melanoma, everybody is talking about targets. None of them has panned out. There’s a lot of hope, looking at LAG3, TIGIT, and other immune checkpoint markers. When we see some activity in other cancers, we’ll probably apply that to liver cancer as well.

What an exciting program. It was brief, but we covered a lot of material. It’s been very informative.

I’d like to thank you, Dr Frenette, or Catherine, and Dr Morse, Michael, for this very insightful discussion. I’d like to thank our audience for joining us, and I hope they found it useful. Thank you to our host, 

™, for hosting this presentation on precision and molecular medicine in liver cancer. Thank you very much for joining us.

Exciting Targets Under Study in HCC

Evolving Approaches for HCC in the Molecular Medicine Era

An in-depth discussion on the targeting of specific molecular pathways with newer, novel therapies to better manage patients with liver cancer.

Evolving Approaches for HCC in the Molecular Medicine Era | Precision Medicine

EP 1: Optimizing Care for Patients With Hepatocellular Carcinoma

EP 2: Hepatocellular Carcinoma Risk Stratification and Therapy

EP 3: Testing for Targetable Mutations in Hepatocellular Carcinoma

EP 4: Targeting the VEGF Pathway in Hepatocellular Carcinoma

EP 5: HCC: VEGF TKI Data and Treatment Selection

EP 6: PD-1/PD-L1 Antibodies and VEGF Inhibitors in HCC

EP 7: Atezolizumab/Bevacizumab Data in Frontline HCC

EP 8: Atezolizumab/Bevacizumab for HCC: Toxicity Management

EP 9: Choosing Between VEGF TKIs for Frontline HCC Treatment

EP 10: HCC Treatment After Previous I/O Therapy

EP 11: Novel Combinations Under Study for HCC

EP 12: Exciting Targets Under Study in HCC