 Pierre, I want to discuss the origin of the liver disease, hepatitis B, C, and NASH [nonalcoholic steatohepatitis]. How do you navigate the instigating element there as it relates to immune therapy? Because we hear a lot of hepatitis B is favorable, hepatitis C is not, and NASH we don’t know. What are your thoughts?

 This has been a question that people have asked since the very beginning of the modern era of effective therapy for HCC [hepatocellular carcinoma]. I think all of the folks here in this audience, maybe except for Mark, remember back in the day when sorafenib came out there was an impression that even in the early days of SHARP and Asia-Pacific [trials], there was an enhanced response in patients who had viral hepatitis. There were studies, some of which I participated that showed that sorafenib has a weak antiviral activity, and reduces hepatitis C viral load by perhaps a half log. This has continued as TKIs [tyrosine kinase inhibitors] came to the forefront. There seems to be a similar impression, even with IO [immuno-oncology] therapy, that a viral hepatitis as an ideology sets you up for a better outcome than nonviral hepatitis, AKA fatty liver disease. Let’s be real here, that’s the bulk of what nonviral hepatitis is. This issue is at the same time important and also poorly understood. The best attempt to understand this at some granularity came from the folks in Germany, Mathias Heikenwälder [, Prof.Dr,] and his group have this paper in 

, which is very much talked about, and shows that in the setting of fatty liver disease, CD8+, PD1+, T cells seem to drive carcinogenesis in the setting of fatty liver disease, perhaps set you up for a worse outcome by dampening the response to IO therapy. That group has taken this a step further and has analyzed multiple studies including KEYNOTE-240 [NCT02702401], IMbrave150 data [NCT03434379], and has shown what we already knew, which was that the response was less in people who had nonviral hepatitis compared to viral hepatitis. Does that mean that I would make different treatment decisions if I had a patient in front of me who had NAFLD, nonalcoholic fatty liver disease, in terms of deciding what treatment option to give them? Within the parameters that Mark and Anthony have outlined, I would do the same thing, but it’s an important question because fatty liver disease has become the leading cause of liver disease worldwide. Hepatitis C is on the decline because it is universally curable and soon will not be a major issue.

Videos

Expert panelists briefly review how a patient’s hepatitis status may influence treatment selection for advanced hepatocellular carcinoma.

Role of Hepatitis Status in Selecting Therapy for HCC

 Mark, in discussions, this gets advertised as the alternative to ATEZO-BEV [atezolizumab-bevacizumab], which as you discussed previously is your preferred standard, especially for those patients at highest risk of bleeding, but then they excluded those patients. Between this and the secondary endpoint of PFS [progression free survival] was intriguingly negative, disappointing is one way to put it. Mark, how do you take this into your practice?

 You really stuck me with the hard question. We have 2 positive frontline phase 3 studies now, both randomized against sorafenib. And maybe more to come that we’ll probably talk about. Technically, it’s not right to do these cross-trial comparisons because these are different trials, but the truth is we need to pick a regimen and we do what we have to do, which is try to parse signals from the data. This is the same challenge that comes up in other tumors, such as renal cell where you have VEGF/IO [vascular endothelial growth factor/ immuno-oncology] vs IO/IO. These are the sort of trends that come up over and over again. The strength of IO/IO is the tail of the curve and sure enough, the tail of the curve for DURVA-TREME [durvalumab/tremelimumab] is very nice. We don’t have that length of follow-up yet for ATEZO-BEV but I think the response rate is higher than DURVA-TREME if you need a response for someone with a lot of disease where you need to get disease control from the available data. 

 Well, let’s ask that question. When do you decide which patient is the patient you’re going to say, you are not a candidate for ATEZO-BEV but I think you’d be a perfect candidate for DURVA-TREME based on the data that we have today?

 There are truly VEGF-ineligible patients, and these are patients with very recent surgery, heart attacks, and recent bleeds.

 Those are the patients where I think there’s no question DURVA-TREME is a compelling regimen.

I follow the same approach, where ATEZO-BEV continues to be the go-to regimen with DURVA-TREME being the backup option. We’ll see how that regimen performs in real-world as well. That remains to be seen.

 Means we must use it in the real world. How are you using it in your clinic?

, it’s not approved at the moment, so I have not used it in the clinic but again, if it is available, it would be for the patients that have clear VEGF contraindications like was just mentioned.

 One point to add is liver function. As you know, I’m interested in ALBI [albumin/bilirubin] score and the meaning for prognosis and efficacy of the drugs we are using now. These are all post hoc analyses, but as said before, we must pick something to make decisions. That liver function is very important in terms of prognosis, but what about efficacy? And what have you learned or what have you seen in the past? I think for most drugs and combinations, we have seen that the efficacy is higher in patients with a good liver function ALBI [grade 1]. This was specifically pronounced in the IMbrave150 [NCT03434379] study; or ALBI [grade 2], ALBI [grade 2a] patients, the curves were almost overlapping with sorafenib. In contrast, in IB1 patients, the median overall survival was not yet reached for ATEZO-BEV, indicating patients with a good liver function are candidates for ATEZO-BEV. What about patients with a more impaired liver function? In the IMbrave study with prospective data, it indicated that these patients might not derive a big benefit. I think they do have a benefit but compared with sorafenib it’s more moderate. Therefore, the real-world data can help. We have recently published data from 14 centers in Germany where we picked 150 patients treated with ATEZO-BEV after the study was published. It was a mixture of patients in the first line and second line, so we did not observe an effect on the line of treatment. That did not matter but, interestingly, we observed a strong effect of liver function and we called them IMbrave in and out. Those that would have fulfilled the inclusion criteria did well, but those that did not meet the inclusion criteria did not do as good as we have thought. This is specifically true for patients with a more impaired liver function. Some of them tried to be and I think this is a little bit in line with the real-world data we are seeing at the moment from other centers. One was just recently published by David Pinato’s [, MD, PhD,] group in hepatology [Imperial College, London, England]. He observed, moderate median overall survival in this patient population of around six months. It’s all like post hoc retrospective, you need to be extremely cautious. Looking at liver function is valid and we need to understand, are there any regimens which might perform better in these patients, which are the majority of the patients I see in my clinic?

Taking into account respective data from IMbrave150 and HIMALAYA, experts consider how they might select frontline therapy for advanced HCC.

Real-World Selection of Frontline Combination Therapy for Advanced HCC

 Anthony, we heard at ASCO GI [American Society of Clinical Oncology Gastrointestinal Cancers Symposium] the results from HIMALAYA. Can you tell us what the study design was and the results?

 Yes, absolutely. HIMALAYA was intended to examine the concept of I/O–I/O [immuno-oncology] combinations. It looked at the combination of anti–PD-L1 durvalumab with anti-CTL4 tremelimumab. Tremelimumab was given in this study as 1 loading dose based on preclinical evidence that 1 loading dose could find the immune microenvironment, and continuation of durvalumab single agent. This loading dose of tremelimumab was a higher dose than normal: 300 mg compared with the traditional dose of 75 mg.

The study had 3 arms and 2 comparisons. One comparison was this stride regimen, which is 1 dose of tremelimumab and a continuation of durvalumab, compared with sorafenib in the control arm. There was another arm of durvalumab single agent also compared with sorafenib. There was no formal statistical comparison of the double I/O tremelimumab-durvalumab with durvalumab. A fourth arm, which people may see on the design that was closed early, examined a different schedule of the T75+D, meaning tremelimumab 75 mg, 4 doses, and continue with durvalumab alone. That was the design.

The primary end point was overall survival looking for superiority of stride, tremelimumab-durvalumab to sorafenib. That was a secondary end point of noninferiority of durvalumab to sorafenib. The study was positive as designed. The median overall survival for the STRIDE regimen was superior to sorafenib in the range of 16 months for the combo vs around 13 months for sorafenib. That superiority end point was met for the combination. Then for durvalumab vs sorafenib, the noninferiority end point was met as well. This was based on a phase 1b study called Study 22, where the response rate was close to 30%, in the high 20%. In this trial, the response rate with the combination was 20%, and there was no difference in PFS [progression-free survival] with the stride regimen vs sorafenib.

 Even at the median, they may have looked a little worse.

 Understood. Those curves flipped a little.

 Some of the strengths of this trial are that it’s the largest randomized phase 3 study in first-line HCC [hepatocellular carcinoma]. It’s mature with long follow-up, so for the first time we’re seeing 36-month OS [overall survival] rates. At 36 months, the combination was superior to sorafenib. Those curves diverged and stayed separate for a long time. We now have 2 positive phase 3 studies in first line, so those are 2 options to choose from. I know the discussion is going to be, how do we choose? I don’t know if you want me to comment on that.

 Not yet. We’ll get to that. I want to understand a little more, and we’ll get to the choices. This study excluded main portal venous thrombosis, VP4. In your viewpoint, what would the implication be? Because the other studies have not. What are the implications of this exclusion in terms of trying to put this study in perspective?

 That’s a valid question. In the REFLECT study, VP4 was also excluded. IMbrave150 did not exclude these patients. We’ve seen a complete response, so checkpoint inhibitors are effective in this patient population. In HIMALAYA, as you pointed out, these patients excluded the question, what does it mean to understand these data? Patients with VP4 have the worst prognosis, which could lead to the question, does this lead to a longer or medium overall survival than either arm? It doesn’t appear so in the sorafenib arm. That was good for the study. Median overall survival wasn’t very long, not 15.5 months as in the COSMIC study; it’s more like 13 months. It was a positive study in contrast to the CheckMate 459 study, which was negative with similar overall survival data. It doesn’t appear that excluding these patients with a worse prognosis led to a significantly longer survival, which we’ve learned for sorafenib.

The other question is, what does it mean for the bleeding risk? With respect to bevacizumab, we were scared about bleeding from varices, and we recommended endoscopy in patients with portal hypertension. Patients with VP4 tend to have varices, and they’re at risk for bleeding. Can we compare the bleeding risk in the HIMALAYA study with the other studies? Probably not. In HIMALAYA, there were no severe bleeding events in either arm, or in the sorafenib arm, which we’ve observed in the REFLECT study. What does it mean for bleeding? With the dual-checkpoint inhibition, we would not expect bleeding risk anyway. I don’t understand why they were excluded. In the REFLECT study, it was a decision by the Japanese society, as I understand, because they don’t want to treat patients with a TKI [tyrosine kinase inhibitor] if they have a VP4. The specific reason to exclude them here, I don’t know.

Shared insight on combination durvalumab + tremelimumab as frontline therapy for advanced HCC in the context of data from the HIMALAYA trial.

HIMALAYA: Frontline Durvalumab + Tremelimumab in Advanced HCC

 Dr El-Khoueiry, please present patient profile No. 2.

 Our second case is a 63-year-old woman who has chronic active hepatitis C infection. She was recently started on antiviral therapy and presented with elevated liver function tests. That triggered a CT scan of the abdomen, which showed multifocal infiltrative liver lesions in the right lobe. There was an enhancing as well, tumor thrombus in the right posterior portal vein. This is evidence of vascular invasion by the tumor and a separate bland thrombus at the splenic confluence and SMV [superior mesenteric vein] and a nodular surface of the liver consistent with cirrhosis.

The platelet count was low, consistent with portal hypertension at 68,000 per mm3. AFP [alpha-fetoprotein] was 5000 ng/mL, bilirubin 1.3 was mg/dL, albumin was 3.7 g/dL, AST [aspartate aminotransferase] was 76 U/L, ALT [alanine aminotransferase] was 112 IU/L, and hepatitis C RNAs were detectable at 1 million [units transferred]. The patient was diagnosed based on imaging criteria with hepatocellular carcinoma. The patient must have shown arterial phase enhancement on the scans with venous phase washout to meet the standard imaging criteria for diagnosis. The patient has Child-Pugh A disease and an excellent performance status: with ECOG 0.

 It’s a straightforward case. Pierre, how would you treat this patient?

 This patient has Child-Pugh A cirrhosis. She is a prototypical BCLC-C [Barcelona Clinic Liver Cancer stage C] patient, with a tumor invading a large blood vessel. She doesn’t appear to have any significant contraindications to the combination of atezolizumab and bevacizumab, so based on the evidence, that would be our go-to first-line model.

 What would make you reluctant to consider atezolizumab-bevacizumab in such a patient?

 There’s a concern when combining these agents for risk of bleeding, primarily related to portal hypertension. One would be well advised to assess portal hypertension risk in this Child-Pugh A patient. We can do this to some extent noninvasively, although the broader hepatology and medical oncology community has relied on endoscopy to assess for the presence of esophageal or gastric varices. The fact that cross-sectional imaging doesn’t show splenomegaly, abdominal varices, or esophageal varices is reassuring.

Depending on where the tumor is located, a FibroScan may or may not be a good test because if you have a tumor, this could alter the stiffness of the liver and give you erroneous information. That’s a possibility. There are other tests for assessment of fibrosis we could do, but nothing points to limitations in our ability to initiate atezolizumab-bevacizumab. There’s no uncontrolled hypertension, and there doesn’t appear to be any underlying autoimmune disease that would be adversely affected by that combo.

 That’s your choice. Mark, what are the data telling us about atezolizumab-bevacizumab, with IMbrave150?

 This is the preferred first-line systemic regimen for anybody who’s eligible for it. It was studied in the randomized study of IMbrave150 vs sorafenib, and this was a positive study regarding every end point that we care most about: OS [overall survival], PFS [progression-free survival], ORR [overall response rate], and quality of life. You need a reason not to use it in first line. Pierre listed several of them, and in this case it’s the appropriate choice.

Centering discussion on a patient with advanced-stage HCC, expert panelists discuss the frontline use of atezolizumab + bevacizumab as a standard of care.

Patient Profile 2: Frontline Therapy for Advanced HCC

 Anthony, here at ASCO [American Society of Clinical Oncology Annual Meeting], everything is about biomarkers. Every disease is moving into the biomarker space to select therapy, to select patients for the right therapy. What’s going on in HCC [hepatocellular carcinoma]?

 A lot in the research realm, but nothing that impacts clinical care. That’s the short answer.

 Are there any promising biomarkers that you would bet on? Is there 1 I want to integrate in my next-phase, randomized clinical trial?

 A lot of drug development in HCC is immunotherapy based. Finding predictive markers in immunotherapy has been a challenge across diseases in general. If we look at PD-1 and PD-L1 expression across recent studies in HCC, including phase 3, you notice that the responses are occurring despite PD-L1 being positive or negative, but the PD-L1–positive patients have a slightly higher chance of responding. It could be an enrichment strategy as part of an overall strategy, but it’s not a predictive marker. There are data evolving about CD8 and CD3 T-cell infiltration associated with outcome, like in other tumor types, if the right cells are in the tumor microenvironment.

 Correct. They tend to do better. There are some immune or inflammatory signatures picked up that consistently show association with multiple end points, including response and survival, including from CheckMate 459, the largest study to report biomarkers. Some of these immune signatures could be taken forward and validated a bit more and may become a part of selection strategy.

 We’re getting closer, but we’re not there yet.

 Closing this margin. Pierre, moving to the earlier stages, to screening and early detection, what’s happening? We keep hearing about platforms that are being looked at, with unmethylated and methylated DNA biomarkers. They look promising compared with other alpha-fetoproteins [AFPs]. What’s moving in that front?

 A lot of these platforms focus on finding suitable alternatives to our already-established screening methods. But in fact, the focus should be how we can augment our screening strategies to make screening more effective and higher yield. I hope the folks who already have products on the market are interested in expanding the use, and maybe take this to heart and start developing products that fill that bill. We don’t have any evidence, although studies—some of which we’re participating in—are ongoing.

These platforms seem to have a sensitivity and specificity that approach or slightly improve upon our established screening criteria for at-risk groups—basically, ultrasound imaging or in people who do this ultrasound combined with alpha-fetoprotein. There are limited studies looking at other markers, which over time have been used to enhance screening efficacy, including AFP-L3 [third electrophoretic form of lentil lectin-reactive AFP], DCP [des-γ-carboxy prothrombin], and things like that. It doesn’t look like that adds a tremendous amount in Western patients. In Eastern and Asian patients, for some reason these tend to be much higher yielding than in our Western patients, probably related to disease etiology and other factors.

 We’re not there, but there’s quite a bit of movement in that space.

 We need something. One paper about the underlying causes of HCC, and NAFLD [nonalcoholic fatty liver disease] was 1 of the fastest-growing reasons for HCC. For this population, we need a screening strategy. We see more patients with obesity and diabetes, which are difficult for ultrasound and to scan the liver completely. We need new screening tools that are low risk.

A brief review of predictive and prognostic biomarker needs in hepatocellular carcinoma, followed by discussion on screening and detection.

Management of HCC: Advances in Biomarkers and Screening

 Anthony, where are you seeing systemic therapy moving into the earlier stages, specifically the intermediate stages?

 I’d like to highlight that the intermediate stage is a very broad group of patients.

 You can go from 1 large lesion with 3 small ones to patients who are riddled with 10 or 15 small lesions in the liver across all lobes. These aren’t equivalent patients, and most staging systems are trying to differentiate these patients. There’s a subgroup of patients with liver-limited disease that’s still BCLC-B [Barcelona Clinic Liver Cancer stage B], who have extensive multifocal or infiltrative disease that’s high risk and should probably go straight to systemic therapy. This is 1 nuance being made in BCLC staging, but it’s in others as well.

Another point is combination therapies for the intermediate stage, which is something we’ve always been interested in doing. Everyone thinks, “We have 2 effective modalities, so why don’t we put them together?” In a nutshell, attempting this approach with TKIs [tyrosine kinase inhibitors] has largely failed. The largest body of data is with TACE [transarterial chemoembolization] or Y-90 [yittrium-90] with sorafenib, and these trials are largely negative. Now there are attempts to combine immunotherapy agents, whether single agent or in combination with TACE or radioembolization. The larger trials are ongoing. There are smaller reports of phase 1/2 studies showing the feasibility of this approach, showing some biomarker changes that validate proof of concept. That is an immune-modulatory effect of the liver-directed therapy. That’s nice and it’s evolving, but we must await the phase 3 data.

The last category compares liver-directed therapy with systemic therapy. This is the newest wave of trials, and it’s looking at patients with classic BCLC-B and asking if liver-directed therapy is still superior to systemic therapy.

 There’s quite a bit of activity. That’s the trend we’re seeing more as our systemic therapy becomes more effective. It makes sense to move it to earlier stages. For patients who aren’t suitable, if you think locoregional therapy may be their best option but unfortunately because of comorbidities, they may not be eligible, Arndt, what’s your standard? What do you do?

 We try to include these patients in clinical trials because we must redefine the intermediate stage and decide who’s a candidate for local and systemic therapies. In respect to this combination of local therapies with immunotherapies, we just presented our data from the IMMUTACE study, which combined TACE with nivolumab in 49 patients. It was a single-arm study with 15 centers in Germany looking at the safety and efficacy. Our primary end point was the overall response rate of 55%, which we counted as clinically meaningful, it was 72%. The largest tumor we treated was 18 cm with TACE. You’d think this isn’t a candidate for TACE, but with the combination with nivolumab, he almost had a complete response. That was remarkable in that case, but overall, the response rate looked very promising. Median overall survival is still early, but we are over 28 months, so that’s a signal that this combination could be meaningful. But I also fully acknowledge that we need to have the phase 3 studies to compare and to address the question of whether sequential therapy might not be the way to go and [that maybe we should] start with 1 treatment and then go to the next.

 Pierre, when we talk about combination, we’re talking mostly about the data with us now, which is single PD-1 or TKI. Now that we’re seeing combination immunotherapy—atezolizumab-bevacizumab, tremelimumab-durvalumab, ipilimumab-nivolumab—is this where we’re moving?

Panelists consider the role of systemic therapy in patients with intermediate-stage HCC who are ineligible for locoregional therapy.

Selection and Use of Systemic Therapy in Intermediate Stage HCC

Anthony El-Khoueiry, MD, associate professor of clinical medicine, Keck School of Medicine, USC Norris Comprehensive Cancer Center, discusses recent combination data in hepatocellular carcinoma (HCC).

There was a lot of discussion surrounding early trials in HCC during the 2018 International Liver Cancer Association Annual Conference, says El-Khoueiry. This includes some trials looking at combination therapy. Most of these trials are looking at anti—PD-1/PD-L1 agents with either tyrosine kinase inhibitors (TKIs) such as sorafenib (Nexavar) and lenvatinib (Lenvima), or with anti-VEGF agents.

One of the trials presented at the conference was the combination of atezolizumab (Tecentriq) with bevacizumab (Avastin). Although it showed a response rate of 60%, it was in a small number of patients, so El-Khoueiry says that this should be interpreted with caution. Further data is awaited. Nonetheless, there is excitement for the possibility for combinations of anti—PD-1/PD-L1 agents with anti-VEGF or TKIs in general, El-Khoueiry concludes.

Anthony El-Khoueiry, MD, associate professor of clinical medicine, Keck School of Medicine, USC Norris Comprehensive Cancer Center, discusses recent combination data in hepatocellular carcinoma.
 

Anthony El-Khoueiry, MD

Articles