: Anjana, one of the most confusing elements, but one that has been ingrained so heavily into how we select patients for therapies is how we score liver function, and we still rely on a system from the 1970s. Give an overview about the different scoring systems and your view. I know you use multiple scoring systems, but how do we make sense out of all these scoring systems when we’re selecting patients for the appropriate therapy?

: Of course. As you mentioned, I think it’s really important to obviously assess liver function when we’re considering any of this. Some of the more common scoring systems, Child-Pugh A, it was from the late ’60s and uses 5 components: total bilirubin, albumin, INR [international normalized ratio], ascites, and hepatic encephalopathy [HE]. We classify patients based on the number of points and severity based on that. Then ALBI [albumin-bilirubin] uses total bilirubin albumin in a logarithmic patient and categorizes to grade 1, 2, and 3 versus A, B, and C in Child-Pugh and kind of takes out the subjectivity of HE and ascites. I don’t tend to use ALBI as much, although of course, I always think about total bilirubin and albumin in the sense of treatments. I tend to use Child-Pugh and MELD [model for end-stage liver disease] more. The underlying thing is, you should look at function as a whole and not rely on 1 or 2 things. What we’re really looking at is, does this person have portal hypertension, what is their ECOG function, what is the extent of their liver cancer? In that sense, using that BCLC [Barcelona-Clinic Liver Cancer] staging system is how you should approach treatment modalities. As long as they have preserved liver function, and Child-Pugh is old, but it still holds true. Those parameters are very important, portal hypertension…it is quite subjective, but it speaks to increased portal hypertension and worsening synthetic function in these patients. As we’re talking about newer treatments, as long as they have preserved liver function, we really have now the option of regression, BCLC stage regression, which is really important. That’s my approach to underlying liver function.

: I agree with Anjana. The Child-Pugh, while old, doesn’t mean that we should just throw it away because it’s old. It has stood the test of time. I agree with Anjana that I don’t necessarily calculate ALBI in all my patients. These newer scoring systems like ALBI allow us to identify Child-Pugh A patients who are closer to transforming to a Child-Pugh B. We’ve seen this with some of the studies that have come out in terms of degradation of liver function with different therapies and that degradation of liver function can differ between different treatment modalities. Considering something like ALBI and whether it’s the calculated ALBI, or taking a look at the individual albumin or bilirubin levels, may identify a patient with Child-Pugh A, who if you “choose the wrong therapy,” could quickly become Child-Pugh B. Therefore, they would no longer be eligible for further therapies, and this becomes particularly important as we have more and more treatment modalities on the table and we’re trying to maximize sequential therapy. The Child-Pugh is a nice base; we’re now developing infrastructure built off of the Child-Pugh that allows us to be more nuanced in terms of our understanding of liver function, particularly longitudinally.

: I agree with your point on that. That is important.

: The only point to add here, unfortunately, it’s really difficult to calculate. With all the scores Philip Johnson, MD, FRCP, has developed, you can hardly use them the clinic. On the other hand, if you look at the difference in survival for ALBI in the Child-Pugh A population, as he published in his initial paper for ALBI 1 and 2, it’s a 10-month survival difference. When you look at the efficacy we are seeing in most clinical trials at the moment, we’re talking about 3, 4, or 6 months at the best. The difference that might come from the liver function is severe, and therefore even if I agree, I do not calculate it in the clinic in every patient there. For clinical trials, it’s very important to be a stratification factor because it has such a strong prognostic and a predictive impact on treatment efficacy.

: Remember that the Child-Pugh score was never designed to assess liver function in patients with cancer. Patients with a lot of cancer, they often have low albumin levels because of their tumor, and if the tumor is taking over a large portion of their liver, they often have high bilirubin because of their tumor. We’ve all seen these patients who are technically Child-Pugh B because of their tumor burden, but their underlying synthetic function is actually OK. It’s something important to keep in mind, and there are patients who are Child-Pugh B who I really think about as being Child-Pugh A with a lot of bad disease. And I treat them more aggressively than I would my standard patients with Child-Pugh B.

: Mark, that’s also important when Child-Pugh B is not a Child-Pugh B. It depends on how you got there. Not only at that one time point, Child-Pugh B7 is very different than a Child-Pugh B9. To your point, what’s causing that Child-Pugh B status? Is it the tumor, is it the cirrhosis, which is super easy for us to talk about, but can be somewhat difficult at times, particularly if they’re newly presenting to you. It’s easier if you’ve been following that patient over time and they’re like, “We’re doing well,” and then all of a sudden developed HCC [hepatocellular carcinoma] and fell off a cliff. You’re thinking, likely cancer. For that patient who had bad liver function coming in, and then developed HCC, that clearly wasn’t the cancer. But it’s sort of the value of video versus a single picture in time.

: Go ahead, Anjana. I’m going to give you the final word on this.

: Longitudinal assessment, that is really important rather than just looking at that 1 point in time, and we all see it where patients all of a sudden decompensate quickly, and it may have nothing to do with their cancer. We also see those who are Child-Pugh B who hold on quite a long time and can get multiple treatments and actually regress to Child-Pugh A. I think that longitudinal component’s also important.

: I just saw a publication on the outcome of patients according to ALBI score in newly diagnosed glioma. When we talk about that Child-Pugh was not developed to predict outcome in cancer, ALBI is now widely used to predict whatever. Maybe this is also something we should keep in mind, that maybe there’s even something outside of this liver field and liver cirrhosis that is also related to the ALBI score. We should be cautious with all these scores; sometimes they are helpful and sometimes not.

: Overall, that continues to overemphasize the importance of a multidisciplinary approach to this disease. This is a disease that’s incredibly complex that relates to the tumor, to the health of the liver, and to the host. It’s very important as we see all these modalities coming together from the systemic, to the locoregional, to optimizing liver function, one cannot but overemphasize the importance of multidisciplinary care.

Videos

Best practices using available scoring systems to evaluate and treat patients with advanced hepatocellular carcinoma.

Scoring/Classifying Patients With Advanced HCC

: Rachna, I know we discussed with you and Mark a little bit the world as it’s evolving around the perioperative setting but there’s quite a bit of interest in the adjuvant setting. We know we’ve been disappointed with an agent that’s mildly active, but what is your view about the evolving world of adjuvant therapies and resectable HCC [hepatocellular carcinoma] specifically?

It’s an area of obviously great interest and I do agree that the natural tendency is always for us to think through perioperative as opposed to just specifically adjuvant. To Mark’s point, a very small percentage of these patients are really amenable to upfront resection, but some of the excitement surrounding it, and there was a study that was presented, the NIVOLVE's trial, which was done in Japan. It looked at 55 patients who had undergone surgical resection or radiofrequency ablation and then received adjuvant nivolumab [Opdivo]. The data again, it was 55 patients, specifically in Japan, you have to put all of that in that context, but the data was certainly provocative. The median recurrence-free survival was a little over 2 years, it was 26 months and that has generated some buzz and excitement. The jury remains out given that that was really just a small study. There’s a lot of ongoing trials right now that are going to hopefully answer this question and everybody had to get into the race, so there’s studies that are specifically looking at patients who undergo complete resection or ablation and are receiving adjuvant IO [immunotherapy] like adjuvant pembrolizumab [Keytruda] as well as combination approaches so looking at bevacizumab [Avastin] and atezolizumab [Tecentriq] in the adjuvant setting, looking at durvalumab [Imfinzi] plus or minus bevacizumab in the adjuvant setting. These are all ongoing and accruing. The fact that they are slowly accruing speaks to the problem of just having patients who can undergo these complete resections and undergo complete ablation or ablation and are eligible for adjuvant treatments. I do think it’s going to take some time. It’s great that they’re global studies because the point is well taken that we can’t just look at one subset and make generalizable statements. It’s an important question. I just don’t know how quickly we’re going to get the readout and if some of the readout is going to be irrelevant by the time we get the readout.

Hopefully, not another storm. The first storm was negative.

Perspectives regarding treatment advances using adjuvant approaches that include immunotherapy to treat early-stage hepatocellular carcinoma.

Adjuvant Treatment Advances in Operable HCC

Talk about something that we can’t shake off our practices. Arndt, this whole concept of using HAI [hepatic arterial infusion] pumps, which has been a concept that’s now 3 decades old. It doesn’t seem to just go away. What are your thoughts?

 It’s really amazing that it does not go away. We are surprised that we see it published in such great journals in recent years. We have to rethink here also a little bit. As we have just learned about SIRT [selective internal radiation therapy] with Y90 [yttrium-90] and that we need to better select our patients where we use these local therapies. The idea of combining different treatment modalities is interesting, and probably we will increase cure rate. It’s also very important that we more clearly define what is our aim. Is it downsizing/downstaging? Or is it really neoadjuvant that we would like to have patients that are resectable, that we would like to increase cure rate? That would be the first point that we need to be more precise [on] here. If I come back to the intra-arterial chemotherapy, which was presented at ASCO [American Society of Clinical Oncology annual meeting], there was one interesting study, which used intra-arterial chemotherapy as kind of [a] downsizing approach. And, as you said before, it’s around for more than 20 years. We have done our studies here as well. If you look at the older studies, the response rate was never so impressive. In our own studies here in Germany, it was around 20%, 30% at best. Now at ASCO in the presentation, it was more than 60%. If you have a treatment with a response rate of more than 60% and hardly any patient who has progressive disease, then it’s an interesting modality for downsizing if you have a patient who is not a candidate for surgery upfront. But then I think it’s important. What is the difference between these patients that are presented today to those that we have treated 20 years ago?For me, the main difference is if you just look at the age, these are young, male, Chinese patients with hepatitis B, and most of them probably do not have liver cirrhosis. If you have a healthy Chinese man with hepatitis B, no liver cirrhosis, you have good performance status, obviously you can perform these kinds of treatment. I was a bit surprised about the 60% response rate. This should be confirmed in other populations. I’m cautiously optimistic whether this kind of treatment will make it into the clinic because progression-free survival is still very short. For advanced patients, it’s not really a method I would invest anything. We have more active treatments with the immuno-oncology-based combination. The only area that could consider it would be really in these younger patients, good performance status, no liver cirrhosis for downsizing, not downstaging. Downsizing, if at all.

Marc, your take as well on the more advanced setting that we’ve seen in HAI [with] FOLFOX [folinic acid, fluorouracil, and oxaliplatin] versus sorafenib [Nexavar], they looked even interesting mildly by a mile. Any thoughts on this?

I agree with what was just said. These studies are provocative. They’re certainly interesting, but we really need replication in a global setting or Western population. The studies that were presented at ASCO, 80% of the patients in one of the studies were enrolled at a single center. I think we need global studies before this can really enter routine clinical practice.

That’s the question. Do we need a global study? Do we need to spend any more resources on studying the question of HAI pump while we have all these modalities? Anjana, I see you have a thought about this.

: I’m very transplant focused. I have seen, as you’ve mentioned, it helps up to a point. But if patients are headed towards transplant, when we’re kind of thinking of these patients on a global sense, it does lead to significant biliary issues and quality of life issues. That’s when we have to pause a little bit and think about long-term. In a subset of patients, it’s quite effective. I don’t disagree. And I agree with you, Tony, I don’t know if we need more study since we have so many more effective treatments, but I would just caution that using HAI in all patients, especially as we’re also moving towards transplanting the subset of cholangiocarcinoma patients as well. It’s just something to keep track of.

: Even more true. Rachna, as a leader in the co-op of a group world and task force, would you even endorse the idea of moving HAI into studying in the cooperative groups?

 My biggest other concern is just the logistics and feasibility of doing a global study. HAI can be done in select centers and to make it a therapeutic approach that has a generalizable applicability. Yes, from a swab like wearing that hat, that would be a really hard study to support and endorse. The logistics of that is really complicated, and that makes it in what are already complicated and sick patients. It’s an exponentially larger problem in that situation.

There’s definitely the complexities of how to manage them. Arndt, I’ll give you the final word on this.

Our patients in the Western world are not really candidates for this kind of treatment because they have too many comorbidities for these natural ability patients, and we have very effective treatments as mentioned before—TACE [transarterial chemoembolization], Y90 for these early tumors. For the bigger tumors, the field for systemic therapy is moving rapidly forward so I would not invest.

Reactions to recent clinical abstracts that support the use of hepatic arterial infusions for chemotherapy delivery in hepatocellular carcinoma.

Hepatic Arterial Infusion Chemotherapy for HCC

 It makes quite a bit of sense. Transitioning to you, Mark, where do you see a lot of our more systemic modalities moving, more so in the early stage, maybe resectable disease, in HCC [hepatocellular carcinoma]? This pre-peri-operative approach, what are we looking at?

: In most prospective analyses that have been done, only about 20% of HCC is truly resectable at presentation. For patients who get resected, the majority of them end up recurring. Early stage HCC is an area where we need some progress. We need to increase the pool of patients who are considered curable, and then for patients who get resection, we need more effective strategies to increase the cure rates. That’s always the goal for the patients we see. There is no other tumor I can think of—we have systemic therapies now that produce response rates of 30% or more—where systemic therapy is not used to cure patients, because that always should be the goal. We presented a small study at ASCO GI [American Society of Clinical Oncology Gastrointestinal Cancers Symposium] that has been accepted now for publication, where we took tumors that were outside of normal criteria for resection. These are tumors either with major vessel involvement or multifocal disease, tumors where normally if they go to surgery, the outcomes are poor because they almost invariably recur. We treated these tumors with neoadjuvant, a combination of a TKI [tyrosine kinase inhibitor], in this case cabozantinib, with nivolumab. Bevacizumab plus atezolizumab would be our standard systemic therapy now, but bevacizumab has a long half-life, which can be problematic in the neoadjuvant setting.

Out of 15 patients, 12 of them ended up getting to resection, and we saw a high pathologic response rate of about 40% in the patients who got that therapy. The majority of patients who had pathologic responses, 4 out of 5 of those patients have not recurred now going on about 2 years. This is a proof of concept study, and there are other studies of neoadjuvant systemic therapy as well. There’s an ongoing study of nivolumab with or without ipilimumab from the University of Texas MD Anderson Cancer Center group that also has found a high pathologic response rate. We need more data, these are just hypothesis-generating studies, but eventually there may be a role for systemic therapy to expand the pool of patients who can go to surgery. At the same time, there are numerous phase 3 studies ongoing of adjuvant PD-1 to address the question, can we improve cure rates in this disease? I’m optimistic that we will.

 Mark, thank you. I want to go back to this, and I want to go to Rachna on this one, that [ipilimumab, nivolumab] data from MD Anderson, which is very interesting, but of course it was a highly selective single-arm study. I know your group under your leadership with SWOG [Cancer Research Network] has been building the next study. Perhaps you can touch upon what this looks like?

: These were provocative and intriguing data. Being a single-arm phase 2 study, it was very logical for us to move this into a larger inner group kind of approach. To Mark’s point, the question is simply when you see a high path CR [pathologic complete response] rate, which is what [Ahmed] Kaseb, [MD]’s study from MD Anderson demonstrated. The question is, can [nivolumab, ipilimumab] be safely given, and can it improve outcomes in the preoperative setting? This study is going to be asking that question. SWOG is running the study, but it’s an NCI [National Cancer Institute]-sponsored study. It’s going to be a question of, should we be giving pre-operative [nivolumab, ipilimumab], and what does that do in terms of recurrence-free survival and outcomes? It’s gone through a couple of different iterations, but it’s now moving forward. We’re excited to be able to do this in a prospective randomized fashion. The important points are going to be simply having buy-in and support from our surgical colleagues. They really are excited about this. There’s a real interest in looking at the potential for neoadjuvant therapy. Given that there were intriguing single-arm data, initially the question was, should we be looking at things like path CR? At the end of the day, what we really care about is prospective recurrence-free survival, overall survival. Those are the really important end points we’re following. 

 I’m looking forward to it, this is an exciting study. I want to go back to Ed. Can touch upon this LEGACY trial, which is a retrospective study, but led to the approval of TheraSphere essentially in patients with HCC. This study was a bit interesting as it included patients with BCLC-A [Barcelona Clinic Liver Cancer stage A] and BCLC-C. What is your perspective about the results of this study, and is it going to change anything in our practice?

: I was pleasantly surprised that the FDA took the LEGACY study and approved the use of the glass microspheres for treatment of HCC. It’s the only therapy and interventional radiology that we have that has an on-label indication for treating HCC. It includes BCLC-A and BCLC-C; it excludes patients with vascular invasion extrahepatic spread.What it leaves is patients with an ECOG performance status of 1, which qualifies them for the advanced category. What it really does is, it takes the radiation segmentectomy, which is what we’re doing with radioembolization; it’s a surgical principle when surgeons do anatomic resection, so wedge resections. They remove that portion of the liver. This is focused on solitary HCC up to 8 cm in size. It had 88% objective response, 84% complete response rates. We’ve been using it for quite a while, but the on-label indication with the FDA helps practices around the country to treat these lesions either as a bridge to transplantation due to the longer wait times now instituted with the acuity circles, or as a curative intent therapy for unresectable patients. As Mark had mentioned, only 20% qualify for a curative resection. This gives the option for patients who have portal hypertension, low platelets, etc, to have potentially a curative intent procedure. We know that many lesions are beyond the ablation criteria of about 3 cm in size for curative intent, and so larger lesions can also now be treated with that type of mindset.

: We should have a quick word about how things have evolved in terms of planning and directing. There is a more directed approach with TheraSphere because the history of Y-90 [yttrium-90 radioembolization] hasn’t always been as focused and directed. It appears that we’re focusing more on these segmental or lobar foci. That’s an important message to bring, that this is a changing modality. The microspheres or the spheres themselves haven’t changed much, but how we use it in today’s world has changed and probably made it more effective and safer.

: You sound like you were an interventional radiologist in your past life. It has been an iterative process, I would say over the last decade, where initially the radioisotopes were delivered in a lobar fashion. Potentially you had higher rates of nontarget damage that would accelerate fibrosis in the liver, especially in the nontumoral parenchyma. It’s changed in our techniques in the last several years to try to self-select just the tumor and the surrounding parenchyma, creating a margin so that you spare background parenchyma. This was shown in the LEGACY study as well, with a 5.6% SAE [serious adverse event rate]. It was shown to be quite safe, despite injecting high radiation doses, which is just internal radiation. It’s embedded into the spheres, and so wherever those spheres deposit, that’s where the radiation is. The radiation radius is only about 3 mm in range. Wherever it’s deposited, that’s where the radiation stays. It doesn’t travel outside of that, you don’t get huge rays of radiation that damage collateral surrounding the area, but wherever that’s deposited, that’s where the radiation is. That leads to safer procedures and also less nontarget damage, and it’s more focused on the target too.

: This is exciting and certainly moving us forward.

Reactions to treatment gaps in the neoadjuvant space for hepatocellular carcinoma and implications for incorporating data supporting newer, novel systemic therapies for appropriate patients.

Neoadjuvant Therapy to Downstage HCC

: Rachna, I want to go to you on this. We’ve seen in the past quite a bit of activity with sorafenib, which is mildly active at best, and around TACE [transarterial chemoembolization] and around surgery, and overall, we’ve seen more failures than successes. As systemic therapy continues to evolve in the more advanced setting, do you see any role for systemic therapy, peri-TACE, or peri locoregional therapy, whatever that locoregional therapy could be?

: There are a number of trials ongoing to investigate these questions. A multidisciplinary conversation that involves looking at locoregional therapy in conjunction with systemic therapy absolutely needs to be part of this conversation now. Referral patterns and such are very different, obviously, across the country and across the globe. Sometimes people come in via hepatology to interventional radiology, and then by the time the medical oncologists see them, they’re Child-Pugh C score, and there’s not a whole lot we can offer. That’s why it’s really important we recognize this collaborative hand-in-hand approach because now to your point, Tony, we have systemic therapies, so it’s a world post-sorafenib. That’s an important question, and the trials that are ongoing will hopefully give us prospective definitive data in this area. Right now, we’re talking, we think that makes sense, but we don’t really know yet.

To your point Ed, about radioembolization and chemoembolization, a lot of the studies right now are looking specifically at chemoembolization, while there’s a shift, I would say at least in the United States, more toward a TARE [transarterial radioembolization] approach. It’s a little frustrating that I don’t know we’re going to get all of the answers we need simultaneously. But I absolutely think there’s going to be a role for that kind of combinatorial approach.

University of Arizona’s Rachna Shroff, MD, describes the role of systemic therapy with locoregional treatment for early stage hepatocellular carcinoma and highlights treatment strategies of interest that remain under investigation.

Advances in Locoregional Therapy for Early Stage HCC

 We’re going move on. We have a diagnosis or suspicion of diagnosis. Ed, I want to move on to you to discuss what you see and how you’re part of this multidisciplinary management of HCC [hepatocellular carcinoma]. It’s been evolving. We’re seeing the locoregional therapies, not necessarily surgical or transplant but the more interventional radiology-related locoregional therapies. Where do we see them fit in the landscape in the evolving landscape of HCC?

 I agree with everything that everyone has said, but we do have to take into account—this goes to your question as well, in terms of what staging we’re dealing with—when we’re in the context of early stage, I would caution the role of biopsy in that situation, especially in the role of transplant. For instance, with mixed histologies, let’s say it’s a solitary 2 cm with presumed HCC. You could get a mixed histology back, and that could hurt the patient in terms of transplant listing. Not to mention the potential for seeding, etc, which could lead to a drop metastasis in the future. I’m absolutely on board. We need a repository with biopsies, especially in academic centers. But in advanced-stage HCC, that’s probably more appropriate, especially because we’re getting more targeted therapies; for the early stage, I would caution with that.

In terms of locoregional therapies and multidisciplinary care of patients with HCC, we always manage in that setting. I have joint office hours with our transplant team and oncologists. It depends on the staging system. We use the BCLC [Barcelona Clinic Liver Cancer] staging system. With the early stage, I see our role with locoregional therapies in 2 fashions: to assist our transplant team as a bridge to get patients to transplantation, which is what we call a curative therapy; or as a curative-intent treatment, which can consist of ablation. The recent SURF trial that was presented at ASCO [American Society of Clinical Oncology Annual Meeting] compared ablation vs surgical resection. The median size was about 1.8 cm. It included up to 3 lesions, but 90% of the patients enrolled had solitary HCC.

We all know that Japan has a very good national surveillance system. They catch these lesions quite early, and they follow these patients for up to 6.5 years on average in each arm. They showed no difference in survival. Ablation has always been on par with resection in this early-stage population for lesions about up to 3 cm in size. We’re finding out more, especially with the recent LEGACY trial that was published, that radioembolization has a potential role as a curative therapy for these solitary HCCs. The LEGACY trial had lesions up to 8 cm and had an 80% response rate with modified RECIST. They also had a long duration of response, greater than 6 months, with 76% of the patients treated. We see that as a role, definitely in the early stage. We know that in the intermediate stage, chemoembolization still has been the go-to locoregional therapy for that stage, but we also know that the intermediate stage is a heterogeneous group. Is it multifocal unilobar, or is it multifocal bilobar? Is it 4 lesions or 12 lesions? There have been variations with the Luigi Bolondi and Kinki University School of Medicine criteria as well, which has a limit of 7 cm. Anything beyond that, patients get systemic therapies. If people are following that, we use those as a guide in our practice and adjust accordingly. I’m sure this will open up a host of discussions.

Edward Kim, MD, of Mount Sinai Health System describes where advances in locoregional therapy for early-stage hepatocellular carcinoma fit into the therapeutic landscape.

Current Role of Locoregional Therapy for Early Stage HCC

 Tony, I said that I don’t biopsy for confirmatory diagnosis if somebody has definite HCC [hepatocellular carcinoma]. But once again, these imaging features have been the best and the worst thing that has ever happened in the field of HCC. I see some of my colleagues here shaking their heads. I think they’re shaking it on the worst side of HCC, more than the best side of HCC—in the sense of you will hear for the next hour-plus—because we haven’t had as many advances in the field of HCC as any of us would like. We’ve seen tremendous, targeted advances in other cancers, but in HCC we continue to be dumbfounded, and our clinical advances tend to be much slower with many more negative trials than we should have had. It’s because of our lack of tissue to facilitate research. I’m personally against diagnostic biopsies to confirm a diagnosis. Academic centers and people who do research I think research biopsies to facilitate a better understanding of HCC, particularly as part of clinical trials. Something we need to be more aggressive with in terms of facilitating an understanding of tumor biology and targets and prognosis.

 Overall, I agree with what Amit said, but I’m not sure whether we should not biopsy in HCC for a couple of reasons. First, we’re all experts, and I completely agree with you. If you have a perfect CT patient at risk with cirrhosis and an expert radiologist, you can make the diagnosis with very high accuracy. This isn’t what we see in the community. The majority of our CTs are not usable. We have to repeat many of them. Many of our colleagues in the community aren’t really expert in diagnosing HCC. Overall, I’m not so convinced that we’re really sensitive enough to clearly distinguish HCC and pharyngeal carcinoma, specifically the mixed-type tumor. We’ve seen that since we started sequencing, it took a while.

 fusions, for example—in mixed-type tumors. Personally, the majority we have we would have met because they were patients with cirrhosis, at risk. Because we always biopsy, we identified them as mixed type, pharyngeal, on your hepatocellular carcinoma, can be identified confusion. Therefore, there are a couple of studies coming out… They’re really careful in the United Kingdom; biopsy is required to prescribe throughout medicinal. They had the chance to compare biopsy and radiological criteria. It’s not as perfect as we think. From our own experience, in 5% to 10% of patients, you misdiagnose, or you make the wrong diagnosis. We’re talking about precision medicine, and that’s not accepted anymore. Therefore, we should be more open, keeping in mind that not only experts and expert CTs are performed. There might be some patients for whom we should be cautious. Liver transplantation is something we of course need to discuss. But overall, yes.

 It’s a great point. I’m speaking from an idealistic perspective in the sense of getting a high-quality CT or high-quality MRI. Your point is well taken that not everyone has access to top-tier imaging or top-tier radiologists. I consider myself fortunate that we do have those at our institution. Many of many of us do. I don’t think we should also undersell what’s available in the community. You do see top-tier imagers and imaging modalities. Your point is well taken that if you don’t believe you have access to that, it’s worth accessing some high-quality imaging and radiologists or considering biopsy in those confirmatory cases. Your second point about mixed tumors is also extremely important. I agree with you that we’re seeing more of these. I don’t know if we’re seeing more of these because there’s shifting in epidemiology or because we’re looking. If you don’t look, you’d never find anything. I’ve been surprised at times when we’ve done these biopsies at how many people do have mixed tumors. We’re potentially treating those people differently. In our center, mixed tumors are the biggest question. I don’t think we necessarily even know how to treat mixed tumors. If you’re going to get a mixed tumor, and you’re like, “I’m going to do the same thing,” then why take a look under the hood? It’s the type of thing for which a biopsy is a small portion. I don’t think you’re able to say on biopsy, is this majority cholangiocarcinoma, majority HCC? Basing your decision sometimes on that biopsy can also be difficult. But your points will take in that if you don’t look, you’re not going to find some of these mutations that can be actionable. I have to say, in our experience, the actionable mutations tend to be relatively rare. When you look at this from a population perspective, at least based on the data I’ve seen—and it may be evolving—the risk-benefit ratio justifies biopsy in all.

 This is a great discussion. It’s a moving target, but I agree, we’re moving into a world where we do need to enhance our likelihood of finding targets to go after. Mark, I’m going to give you the last word before we move on to the next subject.

 It’s great that we have the freedom to start therapy based on a radiographic diagnosis, but I don’t think it’s ever wrong to biopsy a patient with advanced-stage disease. Five years ago, the risk of getting the wrong diagnosis was not that big because sorafenib prolongs life by 3 months. Gemcitabine-cisplatin prolongs life by a couple of months. The risk benefit may not have been there. Biopsies are painful. They have complications, but we’re entering a different era. We’ve all seen patients who are cirrhotic, who have classic risk factors, who have a LI-RADS [Liver Imaging Reporting and Data System] 5 scan—and that’s an excellent-quality scan—who have an AFP [alpha fetoprotein] of 5000 ng/mL. You biopsy them, and they’re a mixed subtype with an 

 mutation or fusion, and they respond nicely to targeted therapy. We’ve all seen this if we do this all day. I’m of the opinion that it’s fine to start therapy when you have a radiographic diagnosis. If there’s something about the tumor that makes you a little uncomfortable—a slightly high CA [cancer antigen] 19-9, for example—then I’m on the side that it’s worth biopsying.

A debate regarding the current role of biopsy to confirm a diagnosis of early-stage hepatocellular carcinoma.

Diagnosing Early-Stage HCC via Biopsy

: Let’s follow through on this idea. Now we have a patient for whom we have a high suspicion they have early stage HCC [hepatocellular carcinoma]. Amit, I’m going to go to you. One of the more controversial issues is the role of biopsy, when to and when not to. How do you stage the patients? How do you use the radiology modalities to essentially avoid a biopsy for some patients and ensure accuracy of diagnosis?

: Yes, great question, Tony. Anjana started to bring this up in terms of, once again, that cancer care continuum. Doing screening is only the first step. The key thing is when you find that abnormal screening result, whether that’s an abnormal ultrasound with an observation greater than a centimeter, or an abnormal biomarker, so for example, alpha-fetoprotein, greater than 20 ng/mL is the typical threshold, then you would need to go on to doing timely diagnostic evaluation. Our diagnostic evaluation in HCC is typically done with a 4-phase or multi phase CT scan, or a contrast-enhanced MRI. I say multiphase because it’s important to have a noncontrast phase, an arterial phase, portal venous phase, and a delayed phase. A single-phase CT scan in the ED [emergency department] does not qualify as a diagnostic test for HCC. What you’re looking for is characteristic imaging features, and that can include features such as arterial phase enhancement, portal venous or delayed washout, capsule appearance, and threshold growth, so growth over time. The nice thing is that we used to look for all of those things, and if you had that, we knew you had a very high risk of having HCC, somewhere between 95% and 97% in an at-risk patient.Patients with cirrhosis or with chronic hepatitis B had a very high risk of having HCC if you had these typical characteristics on imaging.

This has now been codified by the LI-RADS [Liver Image Reporting and Data System], which I think many of our academic health systems and community practices are starting to use more in clinical practice.What you do is you take these imaging features, and you can classify somebody from LI-RADS1 to LI-RADS5. LI-RADS1 is definitely not HCC; LI-RADS5 is definite HCC. If somebody is LI-RADS5, definitely HCC, once again, a 95% to 97% positive predictive value for HCC, that is somebody who I do not biopsy for confirmatory diagnosis because the risk of a biopsy outweighs increasing that positive predictive value any higher than that. Now if you have somebody with, for example, a LI-RADS4, which is highly suspicious but not definite, the risk of having HCC in that lesion is somewhere between 75% and 80%.That can be somebody, based on multidisciplinary discussion and how that would impact your immediate treatment decisions, that may be somebody for whom you say, short interval follow-up, or biopsy. In our center, we do biopsy some of those patients, and once again follow others, but it depends on how that’s going to change your immediate clinical management.

The rationale for diagnostic imaging to evaluate and diagnose early stage hepatocellular carcinoma.

Diagnostic Testing for Early Stage HCC

 titled, “Systemic Therapy for Hepatocellular Carcinoma: A Rapidly Evolving Landscape.” My name is Tony Bekaii-Saab. Joining me today in this discussion are my esteemed colleagues and super experts in the field: Edward Kim, MD; Anjana Pillai, MD; Rachna Shroff, MD; Amit Singal, MD; Arndt Vogel, MD; and Marc Yarchoan, MD. Today we are going to discuss a number of topics pertaining to the use of systemic therapy in advanced hepatocellular carcinoma [HCC]. We will discuss the latest research in the field, and the impact of recent clinical trials on making decisions around treatment selection, which has become quite complex. But excitingly, we’re moving the field forward. Let’s get started on the first topic, and this would be on early stage hepatocellular carcinoma. I want to start with, Anjana. We are trying to understand, as this field is moving more and more toward early detection, trying to bring patients earlier into a more curable setting. What is happening in that in that setting?

: Thank you. First as a background, we can all agree that surveillance is important, especially in patients with advanced liver disease and cirrhosis. That has been set aside. The second thing is that guidelines do vary but recommend using abdominal ultrasound with or without AFP [alpha-fetoprotein] every 6 months. The bigger question, what you’re asking, is how do we risk stratify these patients in the meantime? Who is at highest risk versus lowest? The AASLD [American Association for the Study of Liver Disease] recommends surveillance in patients with cirrhosis only. But EASL [European Association for the Study of the Liver], the European society, also recommends it for advanced fibrosis, or F3. The question is what tools should we use to identify these patients who are highest risk? As hepatologists, I think we have different tools we can use. There are scoring systems like FIB-4 [Fibrosis-4], there’s transient elastography, there are HCC risk calculators for viral hepatitis. All of those things can help us identify, in some sense, who are the patients at highest risk, as well. The other aspect is, what modality should we use? As I stated, ultrasound is what the guidelines recommend, but recent studies have shown about 1 in 5 are inadequate, and ultrasound alone is only sensitive in about 50% in identifying early lesions, like T1 and T2 lesions, and adding AFP increases it to somewhere around 60%, but it’s not ideal. The question is, can we use things like MRI or abbreviated MRI in higher-risk patients? Those are the tools that we have right now.

: Yes, where are we going with all this? The Mayo Clinic has been quite involved with some of that, and you and others as well. Where are we moving into the world of molecular testing and circulating tumor DNA, methylating DNA markers? Do you see this ultimately changing the world of early detection?

: I think so, it’s very important. We’ve all been waiting for these liquid biomarkers, so to speak this Holy Grail, to help us identify these subsets of patients. There are lots of studies out there; the issue is many are low quality in evidence, and so they’re not implemented in clinical practice. I think the HCC early detection strategy at the NIH [National Institutes of Health] Early Detection Research Network, they created these huge biorepositories and databases to look at this specific question. These ongoing phase 2 trials will help us answer that. I certainly think that is the future, and also the evolution of radiomics, extracting imaging data and clinical data to predict outcomes, and ideally with a biomarker, are even more important. I think it’s happening. It’s just not readily available right now.

: What are some of the biggest challenges for early detection? Those patients have to be followed through for long periods, at least those with liver disease who are at risk. In your clinic, what are the biggest challenges and the complexities for increasing our level of efficiency in early detection?

: One is referral patterns, making sure these patients with cirrhosis are referred to a gastroenterologist or hepatologist, to make sure they do undergo surveillance. Amit has done a lot of work on this, making sure there’s provider education, patient education. There are limiting factors, again, of using an ultrasound, including patients who are obese, who have ascites; ultrasound is operator dependent. What do you do with inadequate studies? Do you follow up with cross sectional imaging? Do you repeat the ultrasound again, whether it’s 3 months or 6 months?There’s a lot of disparity in how people use surveillance right now. Those are areas that need to evolve a little better. Then there are some insurance hurdles; I would love to get an MRI on everyone, but that’s not going to fly. And what’s considered cost effective? Those are real problems, and again, effective biomarkers; using AFP over time, longitudinally, the only biomarker we readily commonly use, is great. But it doesn’t detect HCC in 30%, 40% of patients. Using some of these other biomarkers should help in detecting, early surveillance.

: Our hope is moving into an easier world of biomarkers that are more accurate in the field. Amit, you were going to say something.

: Yes, Anjana brought up a great point and it’s an important point. Right now, when we take a look at the at-risk cohort, only 1 in 4 people are getting surveillance per guidelines. It’s important, because we talk about all of the nuances of what we can do and what we should do. It’s important to remember that the majority of patients right now are not getting any surveillance. We talk about, should we do an ultrasound, should we do an MRI? Once again, these are nice debates to have, but the key thing is to do something. As Anjana was saying, right now, surveillance is ultrasound-based, which can be quite difficult to do. There are the provider barriers. We’ve conducted some survey studies among providers. These are pressures we all face, we’re being asked to do more and more in a short period of time. It’s not only with the patient, but it’s all the Epic [health record system]-related stuff that we have to do as well during that same period. There are also patient-reported barriers in terms of getting that ultrasound done. There are financial barriers, transportation barriers, logistical barriers, and those are part of the reason this isn’t being done in the majority of patients. The nice thing about biomarkers is not just increasing sensitivity, and improving test accuracy, but making this super easy on patients, because they see that patient in the clinic, and they get that biomarker done the same day. It can dramatically improve effectiveness of early detection in clinical practice.

A brief overview of criteria and tools used to assess, and risk stratify patients with early stage hepatocellular carcinoma.

Tanios S. Bekaii-Saab, MD, FACP, Mayo Clinic

Articles