&trade; Virtual Tumor Board&reg; focused on hepatocellular carcinoma [HCC]. In today&rsquo;s

&trade; Virtual Tumor Board&reg; presentation, my colleagues and I will review 4 clinical cases. We will discuss an individualized approach for the treatment of each patient, and we&rsquo;ll review key trial details that impact our decisions.

I&rsquo;m Ghassan Abou-Alfa, a medical oncologist at Memorial Sloan Kettering Cancer Center in New York, New York, and the chair of the NCI [National Cancer Institute] Hepatobiliary Task Force. Joining me today are Dr Amit Singal, the medical director of the liver tumor program and clinical chief of hepatology at UT [The University of Texas] Southwestern Medical Center in Dallas, Texas; Dr Riad Salem, chief of vascular interventional radiology in the Department of Radiology and a professor of radiology at Northwestern University Feinberg School of Medicine in Chicago, Illinois; and Dr Farshid Dayyani, a medical oncologist at UCI [University of California, Irvine] Medical Center and clinical associate professor of medicine in the Division of Hematology and Oncology at the University of California, Irvine in Irvine, California. Thank you for joining us. Let&rsquo;s get started with our first case.

Thanks, Ghassan. Our first patient we&rsquo;ll discuss today is a 58-year-old man with a history of NASH, nonalcoholic steatohepatitis, who presents for evaluation. He&rsquo;s a nonsmoker. He drinks sociallyabout 5 to 8 alcoholic drinks per week&mdash;and he&rsquo;s active. He golfs 6 times per week. When he sees us, he&rsquo;s believed to have cirrhosis, so he undergoes HCC screening, he receives an ultrasound, he undergoes

, and he&rsquo;s actually found to have a liver lesion on ultrasound. When you take a look at his laboratory results, you can see here he has Child-Pugh A cirrhosis. With his platelet count, he has a little bit of thrombocytopenia with 142,000. Bilirubin is normal at 1.0 mg/dL; albumin is normal at 3.5 g/dL. He has no hepatic encephalopathy nor ascites, and his

He undergoes further diagnostic evaluation with a CT [computed tomography] scan, which confirms that he has a lesion in the right hepatic lobe measuring 3&frac12; cm. No extra hepatic disease. He undergoes biopsy, which confirms that this lesion is HCC, and the background liver disease shows advanced fibrosis, F3, F4. Overall in terms of staging, he&rsquo;s a BCLC [Barcelona Clinic Liver Cancer] stage A. He has good performance status, 0.

So he undergoes liver resection and does quite well. He has negative margins at that time. Of course, he has a high risk of recurrence, so he undergoes surveillance after the resection. He does well at the 3-month scan but then at 6 months, he actually is found to have recurrence. He has a new liver lesion, 2 cm, consistent with HCC.

The patient is rediscussed at tumor board, and at that point the best therapy was believed to be a radiofrequency ablation. Once again he achieves a complete response at that time. He continues to be followed up on routinely and then does well until about 12 months later, at which time he&rsquo;s unfortunately found to have multifocal disease. He&rsquo;s found to have 4 liver lesions. The largest is just over 3 cm. And more worrisome is that he has 2 lesions in the lung that are worrisome for metastatic disease.

Thanks so much, Amit. This is bread-and-butter, but at the same time there are quite a bit of learning thoughts for us. I&rsquo;ll start with you just because the perspective from the hepatology standpoint will be very critical in regard to decision making for a patient like this. The first question I have is, how critical is it for a patient like your case here to be screened for HCC?

Yes, 1 of the things we have is an identifiable patient population. Over 80% to 90% of people who develop HCC in the United States have cirrhosis. These patients are extremely high risk with an annual incidence, somewhere right around 3% to 4%. And we have effective screening tools. If you do an ultrasound and you do an

test, and you do these every 6 months, you significantly increase the likelihood of applying the HCC at an early stage when we have curative options available. And there have been several studies that show that you actually have an association with improved survival. I think this is critical that we actually identify these patients and put them in a surveillance program.

Fair enough. Farshid, with regard to that screening process that we heard of, and both of us are medical oncologists, what is your diagnostic work-up that you&rsquo;d need for a patient like this? Now there is a mass. It smells like liver cancer. What&rsquo;s your diagnostic work-up that you&rsquo;d to see when you see patients?

The first thing we would do is try to make sure we get the high-quality imaging, triphasic imaging with an MRI [magnetic resonance imaging]. Usually based on criteria, there is an enhancement in arterial case, but the LI-RADS [Liver Image Reporting and Data System] score we would do for the diagnosis obviously. A lot of colleagues will also check the

levels. And in certain cases, if it&rsquo;s a trial patient, for example, we would even proceed with biopsy if the patient was not a surgical candidate for a curative intent patient.

I like what you say. If anything, as you already suggested, with a CT scan, we like to favor triphasic or 4-phase CT or an MRI in that regard, but at the same time there&rsquo;s definitely quite a bit of interest and focus not only on the imaging but also on the blood work to see what kind of extent of cirrhosis could be there.

Case 1: 58-Year-Old Man With Newly Diagnosed HCC

Now go back to our case. As Amit suggested, surgery in the beginning, and there was RFA [radiofrequency ablation]. We&rsquo;ll talk about surgery versus RFA, but Riad, what&rsquo;s your perspective? What do the data show with regard to surgery and RFA?

That is obviously a controversial topic. There are several lines of thought when we think about resection versus RFA. Obviously, RFA has the advantage of being much less invasive. It&rsquo;s done faster, and the recovery is simpler. And when you look at resection, it&rsquo;s a much more involved procedure. There&rsquo;s a surgical complication rate. Of course, most of these patients are going to develop some sort of recurrence at some point. I think it depends on the audience you&rsquo;re looking for.

In terms of this case, I like the ablation option when he recurred. This is textbook, a 2-cm lesion. He probably had washout and baseline imaging criteria. And ablation now has become very well tolerated and very well done with reproducible outcomes. I think this person had the right treatment at the right time.

Fair enough. But I will challenge you here, Amit, on 1 thing. The patient had an initial diagnosis of resectable disease and then recurred. There are data in the literature, among which the data that came out of Memorial Sloan Kettering Cancer Center when Yuman Fong was with us, showing that by definition, a recurrence is stage IV disease. Your thoughts?

Yes, I think the recurrence after resection clearly implies that he has a &ldquo;worse tumor biology.&rdquo; I think you have to be aggressive in terms of management at that time. I actually think that when he had recurrence, he had small disease and he underwent ablation. I think the bigger thing is that there are also data for salvage transplant at that time. Honestly, if this patient was seen at our center, we would have listed him for transplant when he had that recurrence with his 2-cm lesion. We would have treated him with locoregional therapy, but we would have listed him for transplant at that time.

He clearly has what&rsquo;s really shown to be bad tumor biology because even within a year, he had metastatic disease. You can argue that transplant in this patient would have been the wrong thing to do. This patient&rsquo;s course, once again, suggested what you&rsquo;re suggesting, that he was heading down that route toward systemic disease.

Fair enough. And along that line, if anything for our colleagues, the key elements over here are to really realize what are the criteria that make a patient amenable for a resection, RFA, or transplant? And these are the 3 curative therapies we can offer. With regard to resection, it really becomes decision making done by the surgeons. Definitely surgeons will be mostly concerned about vascular invasion and metastatic disease.

With regard to RFA, there&rsquo;s limited regard to size. Usually things that are really very small, around 2 cm, are probably amenable to a RFA. We have seen the data showing that 2 entities, being surgery versus RFA, are equivalent in regard to survival outcome, even though sometimes the recurrence-free survival might be a little better on the surgical point. Again, as Riad said, this is a little debatable.

With regard to transplant, we didn&rsquo;t talk much about that, but if anything, the Mezzaferro criteria, the Milan criteria, remain the basic ones that we go by, which is 1 lesion less than 5 cm; 3 lesions, each of them less than 1 cm; no vascular invasion; and no metastatic disease. From there there&rsquo;s a lot of discussion with regard to who can be transplanted and if it&rsquo;s really amenable. This will bring me back to Riad for 1 thing.

I would be curious. Sometimes we hear about patients with great liver functionality, like our patient over here who Amid just mentioned. Despite that, usually the potential for a transplant will be based on liver function with limited disease, but even with the great liver functionality hereand let&rsquo;s assume the criteria are within the Milan criteria&mdash;why people might argue that this is a great patient to be transplanted as well?

This is the kind of patient where, if they have chronic liver disease, ultimately that&rsquo;s going to be the chance for the best long-term survival. Those are patients who are selected by the hepatology team and, of course, validated by the transplant surgical team. I think the best long-term outcome is one where you&rsquo;re dealing with not only the tumor but also the background liver abnormality is probably the best thing.

I did want to add something, Ghassan, to the RFA discussion, but I think it&rsquo;s 1 of the things sometimes we forget. Yes, small lesions like you mentioned are abatable, but the location of the lesion is also an important narrative, and that&rsquo;s why sometimes lesions are very difficult to ablate because they have the bile duct or the veins, for example, or they&rsquo;re difficult to access or difficult to see on their ultrasound. That&rsquo;s where the stage migration concept comes in. If you can&rsquo;t ablate small tumors, then you migrate to another locoregional therapy like an embolization of some sort. But that needs to be discussed, because that&rsquo;s how we&rsquo;re embolizing a lot of small tumors as well. Because they could be ablated, but it&rsquo;s very difficult based on their location. It&rsquo;s just something I wanted to add to that for the audience.

Case 1: Criteria for Surgery vs RFA vs Transplant in HCC

As we heard, Farshid, ultimately and sadly, the patient recurs. The reality is that a lot of patients unfortunately recur. This is the nature of the disease, as Riad just mentioned. So now we have metastatic disease. What are you going to treat them with?

As you know, for decades we had only 1 option in this schedule with sorafenib based on 2 randomized phase III trials, and it was really a graveyard for our medical oncology with 14 negative trials in first line and second line. More recently, lenvatinib for the first time was shown to be noninferior to sorafenib in a randomized clinical trial in the first-line setting based on the REFLECT trial, which was about a 1000-patient study randomized in first line to lenvatinib dosing based on weight, 8 mg or 12 mg, versus sorafenib. It&rsquo;s primary end point was overall survival noninferiority.

But I think more importantly, there were secondary end points, such as a tumor response, progression-free survival, and time to progression. The trial met the primary end point. It was noninferior in terms of overall survival at 13.6 months for lenvatinib versus 12.3 months for sorafenib. But if you look at the pharmacodynamic outcomes, meaning progression-free survival and response rate, you see very significant improvements with lenvatinib. The cohort, their progression-free survival was almost doubled, 7.4 months versus 3.7 months in the sorafenib cohort with a ratio of 0.66. And importantly there was a 24% objective response rate based on resistance criteria of 1.1 versus 9% of the sorafenib cohort. And that translated into a longer time to progression, almost 9 months with lenvatinib versus sorafenib, which was around 3.7 months. Clearly the patient will be able to stay longer on treatment, first-line treatment, and get a meaningful benefit from first-line treatment if you look at the secondary end points for this patient.

I can&rsquo;t thank you enough. I&rsquo;m really impressed that you know all those data, which are, by the way, fascinating data. If anything, I have to say, you clearly like lenvatinib. I can see that. And you know what, I&rsquo;m not questioning you on that, because I like lenvatinib too. It&rsquo;s a great drug. It&rsquo;s an excellent drug. Let&rsquo;s dissect 2 things, however. We go to lenvatinib, and we&rsquo;ll go back to your data, but let&rsquo;s go back to Amit for a second. We spoke a little in the beginning when you presented the case about the liver functionality and Child-Pugh. Let&rsquo;s talk a little about what Child-Pugh is.

Child-Pugh is a way that we can assess liver function. It&rsquo;s based on 3 variables, and it&rsquo;s based on the presence of 2 things that are hepatically compensation. You can have hepatic encephalopathy, and you have ascites. The 3 lab variables are bilirubin, albumin, and INR [international normalized ratio]. Essentially, depending how normal or abnormal those are, you get anywhere from 1 to 3 points. You add them all up, and so that score goes 5 to 15.

But this is a simple way that we can all objectively speak about liver function and communicate with one another. And it&rsquo;s like grade school: A is the best, and C is the worst. When you take a look at all these trials, these trials included people with good liver function, Child-Pugh A. The tough thing about clinical practice is not everyone has Child-Pugh A disease. You then have to extrapolate these data to people with not as healthy livers, Child-Pugh B or Child-Pugh C patients.

I like the way Amit is describing it. It&rsquo;s really the 5 parameters. There&rsquo;s more or less a certain sequence of how they occur; it&rsquo;s not hazard. It&rsquo;s not that somebody starts with ascites or encephalopathy. But usually the albumin, bilirubin start moving in the beginning. Then this will be followed by probably the ascites. Then again, the encephalopathy and the PTR [proton transfer reaction] will be lost. This is really how the sequence is.

Again, bilirubin and albumin, ascites, encephalopathy, PT [proton transfer], INR. More importantly, even though it looks like a grading of A, B, and C, they are not linear. It&rsquo;s not that A occurs at the same rate of B and C. People live on a really long, long, horizontal line of Child-Pugh A, then there&rsquo;s a little of the curve of B, and then there is shoot of C. That&rsquo;s why it&rsquo;s really important to us to really focus on and try to catch patients very early in the game.

And I&rsquo;ll go back to Riad here. There is 1 thing I would like to hear your comments on, to see if you agree with me. What I&rsquo;m really impressed with in the case that Amit presented is that a patient was caught early with regard to the recurrence. Also, he was acknowledged and caught early with regard to the need for systemic therapy. The patient is still Child-Pugh A. Your thoughts on that?

Look, there&rsquo;s no doubt that I as an interventional radiologist and we in the interventional radiology community recognize that first and foremost, all the data generated for advanced systemic therapy for HCC [hepatocellular carcinoma] are based on Child-Pugh A disease. We know that we have to be on the lookout for that. All the therapies that we have applied, the outcomes are based on Child-Pugh A. So we are actively communicating with medical oncology, with hepatology, and with surgery to make sure that in a continuous process, we&rsquo;re looking at liver functions and making sure that we don&rsquo;t overtreat with any type of therapy. Locoregional therapy included, of course.

The communication with the team starts from day 1. It doesn&rsquo;t start after we&rsquo;ve done multiple locoregional therapies, it starts from day 1. We&rsquo;re working together, and there are some people who decompensate after 1 locoregional therapy; there are some who can tolerate 5, 6, or 7 of them. It depends, but we are actively involved in that, and we know that. We want to make sure that patients are able to benefit from all the treatments, the locoregional therapies, and also the systemic therapies.

I like that. This is very important. If anything, it really reemphasizes the need for the multidisciplinary team. As you see, we&rsquo;re from different disciplines, and we still are discussing the case. Because really this is what we do today. It&rsquo;s very important is that you don&rsquo;t hog a patient because of his certain specialty. Patients need to be seen by the different specialists: the hepatologist, the oncologist, the surgeon, the interventional radiologist, and many others as deemed necessary. By all means.

Case 1: Treatment Considerations for Recurrent mHCC

I&rsquo;ll go back now to the case, and we spoke about lenvatinib. Is there anything you don&rsquo;t like about lenvatinib? Adverse effects?

That&rsquo;s a very good question. There&rsquo;s a distinct adverse-effect profile between sorafenib and the lenvatinib. If you look at the REFLECT trial, you will get more grade 3 hand-foot syndrome PPE [palmar-plantar erythrodysesthesia] in the sorafenib arm versus lenvatinib, which is more defined by generalized fatigue, weight loss, hypertension, and proteinuria, reflecting the more potent VEGF receptor inhibition. Clearly, it&rsquo;s not a walk in the park. We have to be aware of the toxicities and actively manage them. There are some distinct features you have to pay attention to with lenvatinib.

No, I totally agree. If anything, I think we all recognize it and are stressed by it. Especially the hand-foot syndrome with sorafenib can be really quite stressful, even though I give credit: some people really know how to master managing it. But nonetheless, the lenvatinib profile is rather more appeasing in regard to management because at the end of the day, the hypertension is easily controlled. At the same time, definitely a little maneuvering in regard to diet and appetite can help a little in regard to these symptoms. But also, last point, I was really very intrigued by the response rates you brought up. Let me ask you: what&rsquo;s modified RECIST?

Modified RECIST really looks at what makes sense, which means the viable part of the tumor. Instead of measuring the entire tumorswhich might be a 90% narcotic&mdash;and show no change in size in the largest diameter, modified RECIST measures the largest diameter of the hypervascular part, or the part that is actually still viable. That becomes biologically much more relevant in the treatment of hepatocellular carcinoma.

If you look at the correct way, the modified RECIST way, then suddenly the response rate goes from mid-20%, about 24%, to 40% or 41%. That puts the activity of both agents much more into perspective.

Riad, back to you 1 point. We proudly are able to talk to you guys now in IR [interventional radiology], and if nothing else, for us in oncology, now we can talk about response. We didn&rsquo;t have response before. We&rsquo;re very proud of that. To be fair, what does it mean to you? We tell you we have a lenvatinib response rate of 41% by modified RECIST. How do you read what we do?

Yes, Ghassan, you&rsquo;re absolutely right. Historically, with interventional radiology with locoregional therapies, we have been the ones to enjoy the high response rates, no doubt about it. But now with the advent of some of the new therapies, clearly you&rsquo;re catching up, and the response rates are very impressive. Now what does that mean?

I think response to a patient makes a lot of sense. It&rsquo;s meaningful, it&rsquo;s certainly something patients want to hear, they want to know what the tumor looks like: is it small or is it dead, etc? Can we cut it out? But frankly, it also extends the narrative to something that&rsquo;s very new. Historically we used to talk about locoregional therapies for downstaging, but now theoretically, as we start to study this, can we start to think about downstaging with systemic agents?

These are the facts. This is what I&rsquo;ve observed based on the literature. I think we&rsquo;re going to have to be careful as to how we measure tumors with modified RECIST and whether those methods are validated and whether those methods are accepted. But there&rsquo;s no doubt that now that narrative is part of the systemic agent, something that has been usually uniformly enjoyed by interventional radiology.

I like that very much. If anything, we&rsquo;ve learned quite a bit here. No. 1 is that early diagnosis and screening are very critical, as Dr Amit Singal mentioned. No. 2 is that understandably, if there&rsquo;s a potential for a curative intentsurgery, RFA [radiofrequency ablation], or transplant&mdash;we are all for it.

Sadly, as we learned from this caseand we see it with many cases&mdash;recurrence is a possibility. If anything, there was a little bit of a debate about the second RFA. Nonetheless, to cross that bridge, we have seen that ultimately patients recurred and now have systemic disease. We spoke a lot about liver functionality and the importance of making sure that we keep patients, or make sure we manage patients the best as possible as a team, at Child-Pugh A score for chronic liver disease, with good liver function, relatively speaking.

Ultimately, when we need to use systemic therapy, thankfully, as Farshid mentioned, we have many choices, among which nowadays are sorafenib and lenvatinib. We heard quite a bit about lenvatinibif anything, the new advantage, being an anti-FGFR, is that it really has added certain perspectives we have not seen before with sorafenib. Among which, for example, are response rates, which as we heard are 26% in regard to RECIST 1.1, up to 41% with modified RECIST. More important is that ultimately, the patients are living with an appropriate extension of life, as we have seen with the SHARP trial or with Asia-Pacific in regard to sorafenib. And now we&rsquo;re seeing it with the REFLECT trial with lenvatinib.

Case 1: Treatment Response Rates & the Role of IR

Our second case is a 67-year-old woman who presented with a new onset of abdominal pain and was referred for further laboratory tests and imaging studies.

was found to be elevated at 540 IU/mL. She had multiphasic contrast MRI [magnetic resonance imaging] of the abdomen, which unfortunately showed the 8-cm encapsulated mass in the left hepatic lobe showing hypervascularity on the arterial phase and washout in the venous phase. She had more staging studies with chest, abdomen, and pelvis imaging, which showed no metastatic disease. Based on the imaging, she was diagnosed with unresectable hepatocellular carcinoma [HCC].

In terms of her underlying liver function, given the portal hypertension, platelet count was 75,000 mm

. She was categorized as Child-Pugh A6 score for chronic liver disease, based on a bilirubin of 1.8 mg/dL; albumin a little bit low, at 2.9 g/dL; and INR [international normalized ratio] of 1.5. The platelet count we talked about was 75,000 mm

. She had no encephalopathy or ascites at the presentation.

With unresectable disease, she underwent trans-arterial chemoembolization [TACE] X2, and follow-up imaging at 1 month unfortunately showed progressive disease with a rising

of 720 IU/mL. Based on that and based on the data we discussed from the REFLECT trial, she was started on lenvatinib, 12 mg once daily, with stable disease at 3 months. Ultimately at 9 months the MRI showed now progression in the liver, a new adrenal lesion, extrahepatic disease, and

Thanks so much, Farshid. If anything, this case is quite fascinating, especially because from the beginningand I think you mentioned this, if I understood from you&mdash;the tumor was deemed unresectable because of vascular involvement. Ultimately the dissection of the data reflected Child-Pugh A, despite being on the edge with albumin, was not our favorite. Bilirubin was a little high, and a lot of it got embolized. Riad, for you, after embolization that many times, what do you think?

Well, a couple of things. I just wanted to highlight again, these are solitary lesions, large, and technically BCLCA [Barcelona Clinic Liver Cancer stage A], right? All solitary tumors are BCLCA, and this is another example of the state of migration concept when you can&rsquo;t fit the recommendation at that time and then you migrate to the next treatment. In this case, locoregional therapy; in this case, chemoembolization.

There&rsquo;s always a lot of talk about, how many times can you chemoembolize someone? And the narrative of 2 chemoembolizations is really something that has stood the test of time. I actually think that&rsquo;s reasonable. I think there are some data out there that show that you sometimes need the second locoregional therapy, the second chemoembolization, to augment the first response. You don&rsquo;t declare failure until you do it twice.

As I mentioned before in the other case, you have to be careful how many times you do that because different patients will tolerate the number of chemoembolizations differently. This is a multidisciplinary team, so we review the follow-ups in tumor board. There&rsquo;s no such thing as the patients just being seen in IR [interventional radiology] and disappearing for 3 months, disappearing from hepatology, disappearing from oncology. Everybody is reviewed, and we make sure we follow their liver function tests. In this case, I think there were 2 chemoembolizations. The left side is very reasonable, and there was some response here. I assume this was after the second chemoembolization. And I think that was following protocol.

Fair enough. Amit, this setup of the Child-Pugh A was not like, &ldquo;Wow.&rdquo; It was, &ldquo;OK, but on the edge,&rdquo; and with the platelet count of 75,000 mm

, do you think that it was pushed a little more than needed in regard to local therapy?

Yes, I think you&rsquo;re right. This patient once again technically is Child-Pugh A. So they get an extra point for their albumin. As you said, the bilirubin is 1.8 mg/dL. It goes up just slightly, and they get an extra point for that. Albumin is close to the edge. If it goes down a little, they get an extra point, and then they&rsquo;re a solid Child-Pugh B. And you&rsquo;re looking at a completely different story. I think this goes back to Riad&rsquo;s point. It&rsquo;s not as if you hand off the patient and you do multiple chemoembolizations without reassessment. I think you would do the first chemoembolization. You&rsquo;d probably get imaging as well as a reassessment of liver function. And if you find that the liver function is starting to decline, that&rsquo;s another avenue of TACE failure, you may then consider a transition to systemic therapy. I think close monitoring of response as well as liver function as you go through this process is critical.

Of course, what we have here is that we say that the patient underwent 2 chemoembolizations without a lot of detail about what happened between those 2 chemoembolizations. I assume this patient was being followed in a multidisciplinary setting, and I assume the patient&rsquo;s liver function was stable. I think the devil is always in the details, and these patients are always different. It&rsquo;s a matter of how selective you can be. It&rsquo;s the type of thing where if you treat selectively with the chemoembolization, you&rsquo;re able to treat the HCC and have very few off-target affects, and the liver functions naturally remain stable despite being on the edge.

Would you have thought differently, Farshid? It&rsquo;s your case, but would you have loved to start the systemic therapy anytime earlier?

I think we would need a little more information. I agree with Amit on this. Maybe I think in the future, it would be more likely when we have more of the robust data showing that treating the patients earlier will improve outcomes with more output in systemic treatments. I think the answer will be in the future. It&rsquo;s probably much more likely than today on this case.

Yes, if anything, interestingly, for us in the United States we don&rsquo;t have such data, but no doubt our colleagues in Japan have delineated for themselves a certain regimen. How many embolizations? What&rsquo;s the sequence? What&rsquo;s the time between embolization to be sure people are not pushed to the edge before they get into systemic therapy? Anyway, in that case here, the patient ultimately, as you mentioned, got on systemic therapy, got lenvatiniband we spoke plenty about lenvatinib. We like very much, as we said, the noninferiority for sorafenib. We definitely are happy to see that the progression-free survival is almost 3 times higher than for sorafenib. We like the response rate of 26% very much by RECIST 1.1 and 41% by modified RECIST.

But at the end of the day, despite all this good materialand this is an amazing drug, and people can stay quite well on it for a long time&mdash;then people are going to progress. The same as your patient over here.

Case 2: 67-Year-Old Woman With Refractory HCC

What are you going to do with second-line therapy?

Now it gets a little complicated. If you look at the NCCN [National Comprehensive Cancer Network] Guidelines, there are 7 agents listed. We have to pay attention between: is it from the NCCN Guidelines, or is it based on randomized phase III data? That&rsquo;s where the different categories come in, category 1 versus 3. In terms of tyrosine kinase inhibitor [TKI]based or anti-VEGF second-line-plus agents, we have regorafenib, cabozantinib, and ramucirumab for patients with

higher than 400 IU/mL in the second line based on category 1 data. But we also have checkpoint inhibitors, nivolumab and pembrolizumab, at category 2. Actually, sorafenib was also listed for patients who progressed on lenvatinib.

When it comes to choosing the treatment, if you want to be going by data, you can see what the eligibility criteria were in the respective trials and let those guide you. For regorafenib, which was the first TKI that was approved in the second-line setting based on the RESORCE trial, those were only patients who were able to tolerate sorafenib, as defined by at least 400 mg a day for the last 3 or 4 weeks of treatment before progressing. Those patients were randomized to placebo or full-dose regorafenib, 160 mg. And that showed positive outcomes in all primary and secondary end points: overall survival, progression-free survival [PFS], and time to progression.

In other words, what you&rsquo;re telling us here is that, if I&rsquo;m not mistaken, for the approval of the regorafenib prior exposure to sorafenib is necessary, and it should have been well tolerated. In other words, if a patient doesn&rsquo;t get sorafenib beforehand, you are not technically eligible for getting regorafenib, correct?

That&rsquo;s correct. Technically, even if they received it and then had toxicities and then came off, they wouldn&rsquo;t be eligible to go on RESORCE.

Fair enough. What are the other choices besides regorafenib?

The second one is cabozantinib, a combined VEGF/AXL inhibitor, based on the CELESTIAL trial, a large study of second-line-plus. They must have all had received sorafenib, independently of if they progressed or were intolerant. But there were about 20%, 25% of third-line patients on that trial as well. There were not a lot of exclusion criteria in terms of disease volume in the liver portal vein, thrombosis, etc.

The primary end point was overall survival. It was a randomized trial of 60 mg of cabozantinib versus placebo. Again, that was a positive trial with a hazard ratio of 0.76 for overall survival superiority, 10.2 months. But importantly, for the first time, we see with cabozantinib in the second-line settingdespite the inclusion of second- and third-line patients&mdash;with PFS that extends beyond 5 months; 5.1 months in all-comers. If you look at second-line patients only, it actually goes up to 5.5 months. That clearly puts it as a category 1 in second-line plus after sorafenib.

More recently, based on the REACH-2 trial, we have ramucirumab, an anti-VEGF receptor 2 monoclonal antibody, which was tested in patients with an

of 400 IU/mL and greater. Based on the randomized placebo-controlled trialagain, that trial also showed patients with high

overall survival is improved with the addition of ramucirumab versus placebo, with a significant improvement of progression-free survival.

In the second line, we have regorafenib with prior exposure to sorafenib as a condition: if anything, good tolerance and progression of sorafenib. No. 2, we have cabozantinib, which can be used in the third line, as you mentioned. And third, you have ramucirumab but with the requirement of alpha-fetoprotein. Let&rsquo;s talk about the other options in the second line. Let me turn to Amit. Amit, of course we have to talk about immunotherapy. Why don&rsquo;t you start?

Checkpoint inhibitors essentially are a medication that can help use your body&rsquo;s immune system to actually fight the tumor. A tumor basically will express PD-L1 [programmed death-ligand 1] that will then ramp down your immune system. It&rsquo;s a way for the tumors to escape immune surveillance, and these medications are basically able to ramp up your T-cell response and actually fight the tumor.

There have been 2 that have been evaluated in terms of HCC [hepatocellular carcinoma]: nivolumab and pembrolizumab. Both actually showed very durable and deep responses, about 15% to 20%, in large phase II studies. Both have gone on to phase III randomized trials. Unfortunately, both of those phase III trials did not replicate or provide a survival benefit compared with the comparator arm. But the 1 thing we can say is that I think both of these trials, as we&rsquo;ve talked about before, have advanced the field in terms of really showing responsesdeep, durable responses&mdash;with systemic therapy.

Fair enough. In addition to what we heard about the TKIs, we also have 2 checkpoint inhibitors that are approved currently in the United States in the second line, despite limitations based on phase II data that we mentioned in the beginning, being nivolumab and pembrolizumab. This is despite the phase III trials both being negative, nivolumab versus sorafenib in the first line and pembrolizumab versus placebo in the second line.

I&rsquo;m going to go back to that discussion. But there&rsquo;s 1 more important pointby the way, I&rsquo;m sure we&rsquo;re going to have colleagues ask this, or we might even ask the question ourselves. I&rsquo;m going to Riad for this. Recurrence with adrenal metastases. Can&rsquo;t you do anything with IR [interventional radiology] for these adrenal metastases?

You&rsquo;d be surprised what we can do. Sure.

There&rsquo;s no doubt that the manifestation of metastatic disease of HCC is diverse. You can have bone metastases, lung metastases, etc. An adrenal metastasis is a unique form of metastatic disease, and 1 of the ways we can treat some of these sometimes is with an ablative therapy, usually a cryoablative-type therapy if it&rsquo;s accessible.

We&rsquo;ve seen some of that. I think there&rsquo;s been some radiotherapy. Stereotactic body radiation therapy [SBRT] has been applied with these solitary lesions as well. Those tend to be rare, and I think those tend to be the patients where that&rsquo;s the only site of disease. In fact, in some patients, if the time from disease treatment to recurrence is long, they&rsquo;ve been resected because they&rsquo;re better biology patients. In fact, this is an opportunity to think outside the box in the right patient in terms of applying these types of options. There are options for these patients. The reality, however, is they still have metastatic disease, or BCLC [Barcelona Clinic Liver Cancer] stage C, and systemic therapy is still the gold standard. Certainly when you manage patients, they&rsquo;re looking for individualized care, as you mentioned at the beginning of this broadcast. We need to individualize that. There are patients who presented very late who went on to resection, went on to ablation, and lived even longer. Certainly this multidisciplinary tumor board team is what allows you to have this kind of discussion.

That&rsquo;s why I really very much like to work with Dr Salem. We&rsquo;re good friends. But at the same time, I&rsquo;m always fascinated, and it really is important as a multidisciplinary team to understand what are the choices that are on the table, not only from your own specialty but from all other specialties.

I very much like the thinking process that, yes, we can technically do what needs to be done. Of course, we do acknowledge that. But it&rsquo;s fascinating, and look at the attention in regard to biology, that we have to wait for a long period of time to really prove that there&rsquo;s only disease that&rsquo;s still there, and then we can consider a local approach on top of the systemic therapy. But this is really what would require a close and long observation per se.

More important, Dr Salem was talking to us from a rather historical perspective as well. Remember, as Farshid just mentioned to us, we were just at 1 choice a few years ago, and now we&rsquo;re at 7 choices. No doubt things are opening up in regard to more opportunities that we can help patients with. Usually what I tell patients is that it&rsquo;s good to keep everything on the table, and we never say no to anything. At the same time, there is a certain methodical and systemic way to really think about the process exactly as Dr Salem said. Technically he can take care of it, and I&rsquo;m sure he will do an excellent job in regard to an adrenal metastasis. Nonetheless, systemic therapy is the essence over here in regard to management of this patient.

Case 2: Second-Line Therapy Options for HCC

You mentioned to us 7 drugs already between yourself and Amit. How can I choose?

There are different ways to do that obviously, and the more painful way is for every single patient to go back and say, &ldquo;OK, would that patient fit the eligibility criteria for their respective trial? Did the patient benefit for many, many months on sorafenib?&rdquo; Or you try to choose the drug that has the least restriction eligibility, and you&rsquo;re not going to be wrong. Some people might say switching the mechanism of action might be a way to do this from an anti-VEGF tyrosine kinase inhibitor [TKI] to an immuno-oncology [I/O] agent. But I think the data will lead us.

We have 2 negative randomized phase III trials with single-agent I/O versus placebo or sorafenib, as you showed. But we have 3 positive anti-VEGF second-line trials. I don&rsquo;t think the hypothesis is that switching mechanisms of action necessarily will be a guiding principle. I look at the patient, look at the underlying liver function, and look at the performance status. What was the prior treatment? What were the toxicities? Together with the patient we&rsquo;ll make the decision. Most of the time, if I need disease control rather than having the luxury of time to wait, I&rsquo;ll go with the drug that gives me high disease control rate such as cabozantinibfor example, with a 67% range of disease control, whereas initially with an I/O that might drop to 30% or so. So that becomes another way to look at this.

I might challenge you here, but by all means, I totally agree that we might have different perceptions in regard to the disease. If anything, especially if I bring in what Riad just mentioned, I&rsquo;m available. If things are stable, I might really revisit things 1 way or the other. As such, I probably agree with you that response is key.

At the same time, I would like to really revisit the 2 phase III trials in regard to the pembrolizumab versus placebo, and of course the nivolumab versus sorafenib. Yes, they were negative. At the same time, it was not as if they were disastrous. They did not show the primary end point, and I still stand by that the antiPD-1 [programmed cell death protein 1] by itself does not work on the disease. At the same time, the response is very critical over here as it was already seen in the phase II data that we have. As such, I would like to go back to Amit and say, do you think anything was wrong here? For example, for our patients here who got lenvatinib, progressed on lenvatinib, we&rsquo;ll give them pembrolizumab.

I don&rsquo;t think there&rsquo;s anything wrong with it. I agree with you that at the end of the day, when you take a look at the KEYNOTE trial, it&rsquo;s a negative trial. But when you then break down the data and you take a look at this, there was a benefit there. It&rsquo;s just that it wasn&rsquo;t statistically significant because there were 2 coprimary outcomes, and they looked at the data early.

At the end of the day, when you take a look at the absolute survival, there is a survival difference, and there are differences in terms of the responses. I do think that I/O agents play a role in the management of HCC [hepatocellular carcinoma]. I don&rsquo;t think it&rsquo;s wrong to use it in the second line, particularly in somebody who progresses on lenvatinib. Farshid brought up points of how that patient does on lenvatinib. If somebody, for example, doesn&rsquo;t tolerate TKIs well in the first-line setting, I&rsquo;m much more likely to use immunotherapy in the second line. If they go through it quickly, if they basically have disease progression in 3 months, then I&rsquo;m much more likely to use a checkpoint inhibitor in the second line.

I think they play a role. I just don&rsquo;t think there&rsquo;s 1 drug that works for everyone. I think that right now, it would be nice if we had something that would make us smart in terms of choosing. It would be nice if we can actually have molecular pathways that show that if your tumor has this pattern, we would use this drug. We don&rsquo;t have that right now, so we have to use clinical characteristics that got patients into the trial or these other clinical factors, like how quickly you progress on the therapy and drug tolerance.

I totally agree. If anything, at the end of the day the antiPD-1 checkpoint inhibitors in the phase II trials did not really show improvement in survival. But the response data definitely are there and available, and they are already approved for that purpose. Nonetheless, we&rsquo;ll see that our colleagues have more cases to discuss in our tumor board. But if anything, the combination of the checkpoint inhibitors plus TKIs, or even combination of checkpoint inhibitors plus checkpoint inhibitors, is really nowadays a very hot subject. The idea is that if an anti&ndash;PD-1 by itself cannot do itand please do understand that an anti&ndash;PD-1 in HCC is not like an anti&ndash;PD-1 in lung cancer or in melanoma. If anything, a combination of anti&ndash;PD-L1 [programmed death-ligand 1] plus anti-CTLA4 or an anti&ndash;PD-1 plus a tyrosine kinase inhibitor or antiangiogenic would make it. We just have to wait for the data. I&rsquo;m sure we&rsquo;ll have more cases to discuss all those combinations that are coming, for which we are eagerly waiting for the results.

Case 2: Choosing the Right Treatment Option

This case is a 68-year-old man who presented with vague intermittent right-upper-quadrant pain with some weight loss, anorexia, and fatigue over a long period of time. The patient has a significant past medical history of high blood pressure and diabetes, elevated cholesterol, and treated hepatitis C. In terms of social history, there&rsquo;s a remote history of IV [intravenous] drug abuse and moderate alcohol and tobacco use. On physical examination, there&rsquo;s abdominal discomfort and tenderness, ascites, and lower extremity edema.

The ultrasound demonstrates multiple focal liver nodules and enlarged liver. As is appropriate, the patient had a CT [computed tomography] scan that revealed multiple hepatic lesions, ranging from 3.2 cm to 4.5 cm, and metastatic work-up is negative, so there&rsquo;s nothing outside the liver. The liver function in terms of bilirubin is maintained. Albumin is slightly low at 3.0 g/dL, and the platelet count is also low at 102,000 m

. INR [International normalized ratio] is 1.0. When tested for chronic liver disease, this was a Child-Pugh A patient, but by staging is BCLC [Barcelona Clinic Liver Cancer] stage B. This person has a diagnosis of hepatocellular carcinoma with an inoperable condition at this point, based on tumor burden. And the patient underwent 3 cycles of chemoembolization with ultimately a complete response. However, unfortunately, he ultimately suffered worsening ascites and decreasing liver function, again deeming the patient as a noncandidate for future treatment with chemoembolization.

Not good. I give you credit for bringing your case for discussion among all of us. It&rsquo;s not necessarily something that we sometimes see. We don&rsquo;t like to see this. But where do you think really things went wrong? How did we get this patient to get to that level of poor liver function?

I think everything in terms of the work-up, the evaluation, the diagnosis, and the staging was fine. I think the patient was discussed at multidisciplinary tumor board. But I think, as I mentioned before, the reality is that no patient should be managed with locoregional therapy or any of these therapies in isolation of 1 subspecialty. In reality, we do 1 chemoembolization, we assess response, we look at the liver function, and we repeat. And if we&rsquo;re doing well and the patient is tolerating it, we continue to do so. I think, not having images to really look at this, hopefully they treated everything and they did it with a complete response. And presumably after the second chemoembolization liver functions were maintained, which supported the third chemoembolization, which I believe is why they did it.

But one has to remember that these patients have 2 conditions. They have the tumors, they have cancer, but they also have cirrhosis. When that decomposition will occur is unfortunately unpredictable. This is what happened with this patient. Again, I think the recurring theme here is patients are managed in a team, and that includes managing and monitoring the liver functions. It&rsquo;s not in isolation, and it&rsquo;s not 1 subspecialty that manages the patient.

No, I totally agree. Amit, teach us a little here from the liver function perspective. If you can put on your hepatology hat, is there anything that could have been done better or differently to preserve liver function?

Yes. I think there are a couple of things. The good news is that he was actually treated for his hepatitis C before. Actually, you take away that as another insult to the underlying liver function. As we had mentioned, unfortunately, we don&rsquo;t have images to take a look at, but it says multiple hepatic lesions ranging from 3.5 to 4.5 cm. I think it matters where those liver lesions are and exactly how many we&rsquo;re talking about. But I think the thing we do realize is that if you have extensive multifocal disease, even if it&rsquo;s liver localized, some of those patients may be better treated with systemic therapy up front rather than locoregional therapy. I think as a team, we have to be more cognizant that this is a real risk sometimes when we&rsquo;re &ldquo;overly aggressive&rdquo; with locoregional therapy. So I think it&rsquo;s considering systemic therapy sometimes in these patients with multifocal disease earlier.

Then as Riad mentioned, I think it&rsquo;s really following that patient through the course. And so the question is, was this a slowed window where basically after each chemoembolization, we saw the liver function get worse but the patient was still re-treated? Or did they actually have stable disease for the first 2 and then have a big hit after the third one where we saw a sharp decline in the liver function?

The other thing that we do is talk routinely with our interventional radiologists. When they&rsquo;re doing the chemoembolization, if they then say, &ldquo;Look, I can treat the liver lesions but you&rsquo;re going to have a high risk of liver dysfunction because there&rsquo;s going to be a lot of off-target effects,&rdquo; we then will actually say it&rsquo;s probably better not to treat them with the chemoembolization even though they&rsquo;re already in there. We say it&rsquo;s better to stop and to transition to systemic therapy at that time. I think Riad hit on the important points of constant communication between the full team, not just with the tumor board as you&rsquo;re going through the treatments.

Can I just add 1 other thing, as Amit mentioned? Again, while technically this person is BCLC stage B, and the textbook tumor recommendation is chemoembolization, there&rsquo;s a wide spectrum of what BCLC stage B looks like. It&rsquo;s amazing that you can have actually very limited disease and be just outside transplant criteria and be BCLC stage Bwhere you can do a couple of selective chemoembolizations&mdash;all the way to multifocal bilobar disease, which is also multinodular BCLC stage B. There&rsquo;s a spectrum there that people don&rsquo;t recognize. Based on the data I&rsquo;m seeing, in combination with all the systemic agents that are available, early stage B patients are still in the wheelhouse of locoregional therapies. But later stage B situations are clearly things that we need to be careful of, because as Amit mentioned, for me to treat the entire liver, I need much more than 2 chemoembolizations; hence you can go on to systemic therapy.

When you look at the data for locoregional therapy, they really are more so on the early stage B patients than the midstage B patients.

So really, there are very limited data for locoregional therapy on the stage B patients who are more like a stage C patient. It&rsquo;s just like the Child-Pugh class, when you said it&rsquo;s a spectrum. And we&rsquo;re finding the same thing is true for the BCLC staging.

Case 3: 68-Year-Old Man With Borderline Resectable HCC

I&rsquo;m going to come back to some points with Riad. But before I go there, there&rsquo;s 1 important point, and this is a teaching point. If anything, Farshid, I&rsquo;m asking you simply because you were the 1 who presented the data. Remind me, with lenvatinib, when they evaluated it as part of the REFLECT study, there was some comment with regard to how much of the vascular invasion there is. Tell us about what are, say, V4, V3, V2.

If you just look at the top line, you might think the REFLECT study excluded any patient with portal vein thrombosis. But that was not the case. It was really just major portal vein, which is a VP4, and then as you go into the liver and they branch more and more, it becomes a VP3, VP2, or VP1. And the smaller ones were all allowed in that trial. It was just major portal vein thrombosis, VP4, that was excluded in the REFLECT trial.

Thanks for reminding us. It&rsquo;s very important to remember that it&rsquo;s not an absolute criterion. As exactly Farshid is mentioning, patients with a VP4 are going to sadly be sick to begin with, so they might not be eligible for therapy anyway. But with this, let&rsquo;s go back to Riad. You already mentioned that&rsquo;s where you would trigger the question. You mentioned trying to take care of the whole liver. What other methodology could we have used in that setting beside chemoembolization?

This is a prime case in which we can bring up the discussion of radioembolization as an alternative to chemoembolization. Because the reality is 1 of the advantages of radioembolization, which is methodologically very similaryou place a catheter in the artery, and you inject something, in this case you just inject these radioactive beads&mdash;is that this therapy has been shown to be much more nimble in terms of what it can do. It] not only treats the tumor but also gives you a biologic test of time and hypertrophies the contralateral side, so it brings in a narrative of possible resection to patients with whom you might not have had that discussion up front, such as early stage B patients and challenging stage A patients, for example. And so certainly 1 thing to consider in a patient like this in the future is if you have unilobar disease&mdash;and, as we know, cirrhotic patients have small left lobes in general&mdash;you want to hypertrophy them.

One of the techniques we&rsquo;ve been using that has replaced portal vein embolization in our experience is radioembolization because of the advantage. There is a high response rate, you&rsquo;re treating the tumor, and you&rsquo;re hypertrophying the other side. Now we have a large series of patients like this where, as we mentioned before, we&rsquo;re individualizing patient care. All of a sudden you think outside the box, you&rsquo;ve got good tumor biology. OK, well, you&rsquo;re BCLC [Barcelona Clinic Liver Cancer] stage B, but now maybe you&rsquo;re even resectable. It allows you to sort of expand the armamentarium of treating these types of patients, and that&rsquo;s what we&rsquo;ve done with radioembolization.

Teach us a little more about radioembolization. Because in all fairness, some of us don&rsquo;t know much about it. Teach us.

Radioembolization is technically very similar to chemoembolization, where the catheter is placed inside the liver of the artery feeding the tumor and, because tumors are hypervascular, whatever you&rsquo;re injecting accumulates inside the tumor. In this case, you are injecting micron-size radioactive particles embedded with yttrium, with a half-life of about 2&frac12; days or so, so you have about 14 days of activity. And it functions as a low-dose rate brachytherapy-type methodology, but most importantly, Ghassan, it&rsquo;s not an embolization procedure. The issue with, say, bland embolization or chemoembolization is that you block the vessel. Blocking the vessel causes pain, ischemia to the liver. That&rsquo;s how you have postembolization syndrome.

With radioembolization, you&rsquo;re just injecting these beads. They don&rsquo;t change the blood flow, and as a result, patients go home the same day. What you&rsquo;ve done now is theoretically gotten something with a high tumor response rate on an outpatient basis without the issue of occluding vessels, which can happen after multiple chemoembolizations. It really is, in my opinion, the next-generation type of embolotherapy. When done well we can get very, very good results with this technique.

I totally agree. That&rsquo;s great to hear. Amit, I will ask you as a hepatologist. This is definitely a patient who really needs your care. What do you advise our colleagues, especially in oncology, on how to manage or maximize their liver function? What does this patient need from the hepatologist&rsquo;s standpoint?

As we&rsquo;ve discussed, these patients are really best treated in a multidisciplinary setting where the hepatologist is brought in not late in the game, after a patient decompensates, but early and follows the patient through that entire course. This patient started with ascites and lower extremity edema, so he should have been on diuretics at that time and once again monitored through the therapy.

Actually, at this point, because this patient has been treated and actually had a complete response in terms of tumor but has worse liver dysfunction, he can actually be considered as downstaged and potentially listed for transplant. Once again, he&rsquo;ll have to continue having a response and stay within criteria. But this is a disease within a disease, as Riad brought up before. Now that he&rsquo;s had a good response to the tumor, it really is taking care of that liver dysfunction, and transplant is a cure.

Besides the diuretics, if needed, for ascites, lower extremity edema, and sometimes even testicular scrotal edema, what else do you give?

If, for example, he has hepatic encephalopathy, he may need lactulose, he may need rifaximin. You&rsquo;re going to have dietary recommendations in terms of maintaining a good protein diet, low-salt diet. Otherwise he&rsquo;s going to need varices surveillance. The other thing these patients can have is portal hypertension. They can have large dilated veins in the esophagus that can be prone to bleeding, so he&rsquo;ll need to be followed by upper endoscopies as well.

It depends on if he has the presence of these varices. If he has varices present, then you can either band those varices and treat them mechanically or you can start them on beta-blockers. But we don&rsquo;t routinely start beta-blockers outside the presence of varices.

Well, it&rsquo;s interesting becauseand I don&rsquo;t know if it&rsquo;s correct or wrong&mdash;I do understand that some physicians, either hepatologists or oncologists, might put the patient on a prophylactic beta-blocker. What&rsquo;s the value of that? What are the data?

Outside them having varices, starting somebody empirically on a beta-blocker doesn&rsquo;t actually prevent or delay the onset of varices. All you do is you get the adverse effects from a beta-blocker: increased fatigue, etc, without any of the benefit.

Does it help at all in regard to the ascites?

You&rsquo;re decreasing portal hypertension. You may have some benefit in ascites, but it&rsquo;s not the primary management of ascites.

Case 3: Radioembolization, Alternative To Chemoembolization in HCC

I&rsquo;ll go back to Farshid. Sadly, we can see that this is not a good story. It really makes us wonder if things could have happened relatively earlier rather than pushing for one intervention after the other, watching the liver functioning worsening. What would have been your ideal approach with this situation with regard to systemic therapy intervention? What would you have done?

In this period of multidisciplinary clinics, if that patient had presented to our institution today, we would have tried to harness the benefits of all liver-directed therapies, systemic therapy, by putting him on the trial where we do locoregional treatment and combining it with deepening the response by giving an anti-VEGF agent and making it durable by giving a checkpoint inhibitor. I think this would have been my ideal before the patient progressed to that stage.

The other avenue that&rsquo;s opening now with more effective treatment, systemic treatments that actually give us a response, is now something like lenvatinib that gives us 40% response by modified RECIST. If you look at the waterfall plot, many more patients will have tumor shrinkage that is reflected in the RECIST criteria. We would use this as an adjunct or as a systemic treatment to try to cytoreduce this patient, downstage to see whether this patient might become a transplant candidate, for example. The data are not here. We need to generate the data. We have to see whether it will hold up to what we know with locoregional treatment and downstaging and going to transplant. But certainly, now we have the tools to start asking those questions and to try to more aggressively manage these patients.

What we heard so far is that, sadly, this patient who Riad has presented came with a rather relatively advanced cirrhotic component and was treated with embolization on more than 1 occasion while sadly the liver functionality continued to worsen. We&rsquo;re not necessarily saying this could have happened, but naturally this is really what occurred at the end of the day. Already as we can see on the table, there are other options that could have been there, including what Riad mentioned: going to radioembolization that probably could have controlled disease in a more expansive way than the localized chemoembolization. We heard about potential for certain further intervention, now that the responses are available, starting with lenvatinib but also with the checkpoint inhibitors, as we mentioned, pembrolizumab and nivolumab.

In addition, we have data that suggested that we could also bring in a great response. Among which, as we all have seen, there was a press releasewe don&rsquo;t know the data yet&mdash;of the combination of atezolizumab plus bevacizumab. We have also the data that were reported on lenvatinib plus pembrolizumab. Also we have the durvalumab plus tremelimumab, the anti&shy;CTLA-4 plus PD-L1 [programmed death-ligand 1] inhibitor strategy. There&rsquo;s quite a bit going on, as Farshid already mentioned. But 1 thing that all 3 of my colleagues have brought up&mdash;and I have to admit that I don&rsquo;t have full understanding of this and I&rsquo;m going to go back to Riad on this&mdash;is downstaging. What&rsquo;s that?

is a term that&rsquo;s used in the transplant context. While we know that patients who present within Milan criteria for liver transplantation have a certain 5-year survival, is it possible that if you take somebody outside those criteria and shrink them and make them within those criteria, and then demonstrate that the biology is there to support that, can you impart the same type of benefit? More recently that concept has been expanded to the surgical side in cirrhotics. Can I downstage somebody who&rsquo;s initially inoperable for whatever reasonsize, for example, or location of the tumor&mdash;into somebody who&rsquo;s more operable?

The idea here is that you take a patient who&rsquo;s otherwise not transplantable, not resectable, and you downstage them; that&rsquo;s the term. You potentially make them transplantable or resectable to a possible cure by using a locoregional therapy. That&rsquo;s what we have traditionally enjoyed in interventional radiology with the locoregional therapies, where you have this. But with the advent of some of these therapies, like lenvatinib with 40.9% response rate or so by mRECIST [modified RECIST], now that narrative is there in level 1 evidence, can we use mRECIST criteria to support that? I think that narrative is being expanded.

I can see you&rsquo;re ready to give your view. And you know, I&rsquo;m very eager. I was going to ask you because I&rsquo;m not sure I buy this.

Oh, no, I completely agree with Riad, actually. But there are data that show that if you actually get downstaged, you have equivalent survival as those who start with the Milan criteria.

This started from the University of California, San Francisco group. There were actually data from Texas and Oklahoma as well, and now this is national policy.

I think this really is important for everyone who treats HCC [hepatocellular carcinoma] to know, because you started by describing the Milan criteria, which is the traditional criteria that were described by Vincenzo Mazzaferro and has been the standard for years. Downstaging now means that for anyone who has locoregional disease, you have to consider, can I downstage this patient and make them a transplant candidate? This is an important concept because those patients then have 5-year survival that exceeds 60%. This is a cure for their liver disease. This is a cure for their HCC.

Farshid, you mentioned that when we think about systemic therapy options, we can use the tyrosine kinase inhibitors and immunotherapy. But I think it&rsquo;s important that these patients are discussed in a multidisciplinary fashion, because if you treat that patient with locoregional disease with immunotherapy, at this time we don&rsquo;t know if that affects post-transplant outcomesand if so, how it impacts post-transplant outcomes. And so if you treat a patient like this, who has liver dysfunction and locoregional disease, with 1 of the checkpoint inhibitors, you may be impacting their chance of getting a transplant and their post-transplant outcomes. I think that&rsquo;s where early referral and discussion in a transplant setting, a multidisciplinary setting, is critical for these patients.

Case 3: Downstaging and Increased Survival in HCC

This is a 67-year-old Caucasian woman without any known risk for cancer, including cholangiocarcinoma, who was admitted with acute abdominal pain, nausea, and vomiting. The imaging findings on the CT [computed tomography] scan and the MRI [magnetic resonance imaging] of the abdomen showed multiple intrahepatic lesions of undetermined nature and 2 small lesions in the lung, left lower lobe. Biopsy of the hepatic lesions was suggestive of intrahepatic cholangiocarcinoma. I&rsquo;ll start with Amit. What&rsquo;s cholangiocarcinoma?

When you think of liver cancer, there are actually multiple types of liver cancer. As we&rsquo;ve talked about for the first 3 cases, we focused on hepatocellular carcinoma [HCC], the most common type of primary liver cancer, which comes from the cells of the liver. Cholangiocarcinoma is the second most common primary liver cancer, and it comes from the actual bile ducts, typically of the liver.

Fair enough. Farshid, to help our friends, our colleagues, how do you treat it? It&rsquo;s a stage IV disease here, understandably.

Historically, recurrent unresectable metastatic cholangiocarcinoma we treat with systemic treatment. Based on the ABC-02 trial, a combination of cisplatin plus gemcitabine was superior because they followed up with the second-line FOLFOX trial, which showed a survival over gemcitabine alone. That&rsquo;s been our standard for a long time. I think the field is moving. There&rsquo;s more molecular understanding of the biology of cholangiocarcinoma.

We&rsquo;ll come to that, but if anything, this lady should get gemcitabine-cisplatin.

Definitely. She has the correct parameters.

I totally agree. Let me, before we dissect a little further, reference the other systemic therapies. Riad, let&rsquo;s assume there are no 2 small lung lesions over here. The disease is disease in the liver only, 1 mass that is rather large, and there are some satellite lesions around. This is a classic scenario we see in cholangiocarcinoma. Is there anything you can help with here, from the interventional radiology standpoint?

Well, I&rsquo;m glad you asked because the reality isand I&rsquo;ve been waiting for this case to talk about this a little more&mdash;I think 1 of the things we&rsquo;re really lacking is more prospective controlled trials actively combining the local and the systemic therapies. This is true of HCC, and this is true of cholangiocarcinoma. So you&rsquo;re right, this is a very common scenario, that central lesion with the scalloped anterior liver with a couple of lesions. The reality is, if you pick a patient who has good liver function, we have very few adverse events when you do a chemoembolization or radioembolization in that area. And nobody is suggesting that you do that in isolation.

What you would do, however, is hope to augment the results because, at the end of the day, the results you&rsquo;re quoting from the ABC-02 trial aren&rsquo;t that impressive. This is what you mentioned. There are level 1 data, I get it, but the survival is still 11 months or so. One of the nice things would be to try to combine these types of things in a more prospective manner. This is 1 of the disease conditions in which it&rsquo;s completely lacking. Everything is still systemic in terms of treatment options. That&rsquo;s my bias as an interventional radiologist. But again, if you pick the right patient, you know you can control your adverse events and you&rsquo;re not going to deny anybody the standard of care, which is what I think you were alluding to.

Fair enough. I totally agree. Now, Farshid was about to discuss the potential biologic therapies. But let&rsquo;s talk a little about the biology of the disease itself. What do you know about that?

Cholangiocarcinoma is an aggressive tumor. We know that liver cancer overall, both HCC and cholangiocarcinoma, have a relatively poor prognosis. This is a disease for which the 5-year survival remains less than 20%. But I think when I think of HCC and I think of cholangiocarcinoma, I think of cholangiocarcinoma as actually being more aggressive even than HCC. These cancers are typically found at a late stage. We don&rsquo;t have an identifiable target population as much as HCC. There are really no screening recommendations per se. Of these tumors, unfortunately, most of them are found advanced. Even when they are found early and you are able to resect them, there&rsquo;s a high recurrence risk, and the survival is just not as good as it is for many other cancers that we see.

Absolutely. Farshid, I do next-generation sequencing on this patient. What are the things, if any at all, that you would like our colleagues to look for?

I think the list is growing, but as of today, I think

alterations. I think for extra biliary or extra liver disease, we look for HER2 alterations.

For gallbladder, for example. And there are subtypes, such as

V600E mutations, where we have some data with combinations of MEK and RAF inhibitors. These are the ones at the minimum, but if you send for next-generation sequencing, you get obviously a much larger panel, MSI [microsatellite instability] status, for example, as well, for the very small proportion that might be MSI-high.

Let&rsquo;s dissect this a little further, and let&rsquo;s start where you ended. Tell us a little more about MSI-high disease, which ultimately pertains not only to cholangiocarcinoma but to all diseases. But in MSI-high disease, what does it mean? What do you treat with? How common is it?

Microsatellite unstable tumors, MSI-high tumors, have a higher rate of mutation. As you know, pembrolizumab was the first drug that was approved independent of histology for all solid tumors that are MSI-high in the second subsequent line of treatment. The rate of MSI-high disease depends on the site of disease. For liver cancer, it is in the single digits, depending on what series you look at. For colorectal cancer, it may be as high as 15% or 20% for early-stage disease. But there is a much higher probability of response and durable response with checkpoint inhibitors in the MSI-high population. It&rsquo;s definitely worth testing the patients because if they are MSI-high, they would potentially significantly benefit from checkpoint inhibitors.

Fair enough. Again, as Farshid said, this kind of population is rather limited. It will make less than 5% even. It&rsquo;s not something we&rsquo;re going to commonly see, but nonetheless, it&rsquo;s totally vetted and legitimate to check for MSI-high patients. Especially, of course, if the historyfamily history, especially, might suggest that or pertain to it.

But I&rsquo;m definitely more intrigued, and admittedly we are all heavily involved in regard to this work. But the

, if anything, is a mutation that will actually make an alteration with regard to the ability to 2-hydroxyglutarate, which likely is normal in all our Krebs cycle. But we bring in alpha-ketoglutarate, which if anything, is actually an oncogenic protein and as such can cause an unprecedented uncontrolled growth of the tumors. Lately we have a report that looked at ivosidenib, which is an anti-

, versus placebo in second-line setting with a crossover that allowed patients on the placebo to go to the IDH1 inhibitor. It showed actually an improvement in progression-free survival as well as overall survival. These are definitely robust data that at least will bring in a potential for the anti-

Case 4: 67-Year-Old Woman With Intrahepatic Cholangiocarcinoma

Interestingly, another category of drugs that is quite intriguing is the

fusiontargeted agents. Let me ask again 1 more time, Farshid. Tell us a little more about

. Are we looking for a fusion or a mutation? Which one?

I think we are looking more for fusion rather than mutations in cholangiocarcinoma. And the data that we have show that mutations probably don&rsquo;t work as well from other solid tumors than alterations that we look for. We have 2 agents that have been observed in relatively large single-arm trials with high response rates in the 25% to 35% range and, importantly, high disease-control rates in the 75% to 85% range with infigratinib and pemigatinib recently. Those will be validated, I think, in larger phase III trials.

historically seems to be maybe favorable prognostic factor in intrahepatic disease, but I think now it becomes more of a predictive factor, especially as fusions are detected by NGS [next-generation sequencing].

I totally agree. If anything, there are quite a bit of intriguing trials. As you mentioned, the infigratinib and pemigatinib are actually being looked into first line versus gemcitabine and cisplatin as the standard of care, exactly as Farshid already mentioned to us. Interestingly, there is a crossover prospective, and we&rsquo;ll see where these might define or decide where the anti-

fusions will come into play. Nonetheless, there are other agents that really are being looked at, among which are debio 1347, TAS-120, etc. We are trying to understand better where they will take us.

Interestingly, if anything, the message over here is that despite nothing yet being approved by the FDA, at least we have the MSI-high strategy, which is enough suggestion for our patients to make sure that we do the next-generation sequencing for all those patients. That brings up a very important point, and I&rsquo;ll go back to Riad over here. When do we get the next-generation sequencing? Do we wait until the patient needs a systemic therapy, or should we do it from the start? Is this on your mind-set as an interventional radiologist, or would you like to pass on a message to your colleagues in interventional radiology?

I can address the tactical issues that exist with that. Certainly, based on where the lesions are, these are things that we can get up front. There&rsquo;s no doubt that we are in a world in which biopsying up front and actually after treatment looking for mutations is something that we&rsquo;re seeing. Now the oncology practice and interventional radiology have become very biopsy heavy because of the advent of all these agents. For me it seems like it starts certainly up front as a baseline biopsy, but certainly there&rsquo;s no doubt that that continues to increase. As you&rsquo;re switching classes of agents and types of treatments, there&rsquo;s no doubt that the number of biopsies is increasing, and this is something that clearly we&rsquo;re able to do in interventional radiology.

I totally agree with Riad. This is our message and our obligation to talk to all our colleagues. One more time, we revisit the multidisciplinary approach. When we have somebody who is going for a surgery or going for a local therapy, we would like to make sure that a piece of the tissue is removed at surgery, or a biopsy at the time of the local therapy, or even an intervention such as a biliary drain. That&rsquo;s needed to make sure you get tissue, so we can get the diagnosis established for next-generation sequencing. Because obviously this might take a little bit of time, and having these data ahead of time can really save a lot of time for the patients to get on the correct clinical trial and be treated accordinglyin no time, hopefully, because we need to make sure we add more options for our patients in regard to therapy. We will see those drugs really prove their point in regard to treatment.

Case 4: Next-Generation Sequencing

I&rsquo;ll go back to Amit for a second. You brought it up, but let&rsquo;s dissect it a little further. Cholangiocarcinoma plus hepatocellular carcinoma [HCC]?

Yes, we&rsquo;re seeing this more and more. We used to think of HCC and cholangiocarcinoma as primary liver cancers, but we thought of them as 2 distinct diseases. I think more and more we&rsquo;re starting to see that it&rsquo;s really probably a spectrum. As biopsy is starting to be used more, even for HCC, in our clinical practice, we&rsquo;re seeing more of these mixed tumors. I think you see some that are predominantly HCC with a touch of cholangiocarcinoma, and then you see some that are predominantly cholangiocarcinoma with a touch of HCC.

I think it&rsquo;s going to be an interesting time because for those mixed tumors, I don&rsquo;t think we really know exactly what the optimal treatment is. I imagine it really depends on what your predominant tumor is within each of these. But we do see more mixed tumors. At least in our clinical practice, this is why we&rsquo;ve become much more prone to biopsy, even things that may have characteristic imaging.

Fascinating. Let&rsquo;s talk about biopsies. I&rsquo;ll go back to Farshid. For a patient with cholangiocarcinoma, a certain set of mutations progress. Another repeat biopsy?

At the minimum, I would say liquid biopsy. As Riad already alluded to, tumors are dynamic. There are new mutations emerging, resistance mutations, so we do try at a minimumif we can&rsquo;t access the tumor for whatever reason, for safety reasons&mdash;to do a liquid biopsy. Now, these tests take time to come back. This is an aggressive disease, Amit said, so you want to do your homework early on and get early profiling. But upon progression, at a minimum, we try to get a liquid biopsy to see if there are any emerging mutations that might be targetable that were not in the original tumor. If feasible, obviously we would also do a tissue biopsy.

I like what you said. Interestingly, there are dataand I&rsquo;m sure you were aware&mdash;that came out of Massachusetts General Hospital. It was really a very nice piece on patients with

fusions who progressed; it&rsquo;s fascinating. This is where they actually had a statautopsy collection of tissue from liver&mdash;and found out given the different locations of tumor, what are the changes in regard to the genetic mutations or the genetic fusions? It&rsquo;s quite fascinating that really, yes, these changes do occur. We&rsquo;re not necessarily vouching for making sure that we rebiopsy everybody. Ultimately, the science of it still needs to evolve. At the moment, actually, pathologists will tell us that not everybody will have an alteration in regard to their genetic makeup of the tumor. But then we&rsquo;re learning more and more about this.

The fascinating part about it is, is it really something that happened by randomie, it changed from 1 mutation to the other? Or is there a certain phenomenon that would make a tumor stressed enough to really alter to a different mutation? Time will tell, and we&rsquo;ll see how it&rsquo;s going to take us from there. Go back, however, to the

. Some of our colleagues get a little nervous about the anti-

because of the adverse effects. What are the things that we are mostly worried about, and what are your thoughts?

Retinopathy or serious retinopathy is something you need to pay attention to. That means closely working with an ophthalmologist, or eye doctor, referring the patients. And it&rsquo;s manageable. You have to hold it, let it resolve, and then you can actually restart the patients. That&rsquo;s something. You look at their laboratory results, phosphate level. Hyperphosphatemia is a very common adverse effect. With sevelamer and other phosphate binders, you can manage that. There are some skin and nail changes that you look for. There are adverse effects that you have to be aware of, but you have to closely monitor your patient and work very closely with the eye doctors at a minimum to manage these main adverse effects.

I think you named the 3 key ones, which are hyperphosphatemia and the skin, nail changes, and retinal changes. And close monitoring would be key. It&rsquo;s quite fascinating, though. Hypophosphatemia is also noted, but the best understanding we have is that really was physician driven. We were so excited about the hyperphosphatemia, enough that we caused the hypophosphatemia. So we should not treat numbers but rather treat symptoms. This is really where it&rsquo;s very important to get a better understanding about the management of those symptoms.

Hepatocellular Carcinoma