Which Patients With HCC Are Suitable for Treatment With Atezolizumab and Bevacizumab?

Anthony B. El-Khoueiry, MD: Dr Rich, atezolizumab-bevacizumab has been the standard of care. In which patients do you think that continues to be the standard of care? What has been your experience with this combination?

Nicole Rich, MD: Since it has come onto the scene quite spectacularly, it’s been the first-line option, particularly for patients with unresectable HCC [hepatocellular carcinoma], who mainly in our practice would have advanced-stage HCC or BCLC [Barcelona Clinic Liver Cancer] stage C. Although increasingly, we’re talking about the role of systemic therapy in general, including given the high objective response rates with atezolizumab-bevacizumab in patients with intermediate-stage HCC or BCLC B as well. Particularly those beyond up to 7 criteria, or those with more infiltrated tumors that aren’t as amenable to locoregional therapy. Those are the 2. It remains standard of care for advanced-stage patients, BCLC C, as well as some consideration in select Bs. It’s also used in patients who don’t have the contraindications you mentioned. We’ll get to that in a moment.

In our experience, this has been a well-tolerated regimen. The main limitation is the patient’s ability to travel to get their infusion. That would be 1 of the main concerns compared with a TKI [tyrosine kinase inhibitor], where it’s an oral regimen and maybe a little more convenient for the patient. But in general, it’s been well-tolerated. We haven’t seen many adverse events at our center.

Anthony B. El-Khoueiry, MD: Great. To summarize your excellent point, who are the patients with HCC who should get systemic therapy? We use the BCLC staging very commonly, but this may not be something familiar to everyone. Patients with BCLC stage C have either main vascular invasion in the liver, meaning main portal vein invasion, or extrahepatic metastasis. That’s easy. Patients with BCLC B have liver-limited disease that’s usually multifocal, but that’s a very heterogeneous group. It could be 2 lesions; it could be 10 lesions. There could be a difference in the size and distribution of these lesions. Usually, patients who have a heavy burden of liver-limited disease and up to 7 criteria could be 1 thing to follow, or those with very large lesions in the liver. These patients can go to systemic therapy straight away because that’s the highest level of evidence for a survival benefit from any therapy for HCC.

The last subset of patients are those with BCLC B disease who have had liver-directed therapy and are no longer candidates, either because of progression or because they’re not tolerating it well. It’s important to note that this transition from liver-directed to systemic therapy is a critical transition because we don’t want to delay it and allow patients to have deterioration of their liver function and miss the opportunity of systemic therapy. Dr Rich, you alluded to this. Which patients shouldn’t get atezolizumab-bevacizumab? Where do we need more data?

Nicole Rich, MD: There are the contraindications in patients who were excluded from the trial. The main concern is bleeding concerns in cirrhotic patients, given the anti-VEGF that was mentioned before, the bevacizumab component. For this reason, all these patients were up to date on their endoscopy and variceal surveillance. They all had an EGD [esophagogastroduodenoscopy] within 6 months and appropriate treatment of varices, as you mentioned.

Patients with a history of repeated variceal bleeding would be something that would cause you to take a pause. This is especially important in patients with new macrovascular invasion or new portal vein thrombosis, even if they had had an EGD 6 months or a year before. They could have worsening of their portal hypertension in that time as a result of the tumor extension into the portal vein. Bleeding in a cirrhotic patient is an important consideration. Another important consideration is patients with cardiovascular disease that’s severe with recent ischemic events. That’s another population in which you would be worried about the bevacizumab component. The other issue that comes up is the patient’s liver function, which, as a hepatologist, is the area where I typically get involved with our medical oncologist and help comanage these patients. We have been using atezolizumab-bevacizumab in patients who are good Child-Pugh Bs. For example, Child-Pugh B7s in whom it may be because of albumin, not folks with ascites that’s not controlled. Those are the main considerations from my standpoint.

Anthony B. El-Khoueiry, MD: Very good. Dr Gholam, what are concerns about treating Child-Pugh B patients with this combination? What data do we need?

Pierre Gholam, MD: At this point, we remain somewhat limited in our data sets or clinical information that reflects safety and efficacy in the setting of people who receive immunotherapy. We have some data previously that were reported in the setting of nivolumab therapy. We have some data from Korea that were presented at the recent meeting that also reflect some questions regarding mitigating risk of bleeding and whether that might be too high a cost in terms of implementing such a treatment. But the bottom line is we remain somewhat without clear guidance about whether we would be able to safely and effectively treat those patients in that setting. This has been acknowledged by the NCCN [National Comprehensive Cancer Network] and others, where they have specifically allocated additional guidance regarding Child-Pugh score to reflect the paucity or relative abundance of data, especially for TKIs, in that setting.

Anthony B. El-Khoueiry, MD: Great. Dr Rich, I’d like to go back to you about Child-Pugh B patients. Do you want to add anything else?

Nicole Rich, MD: I want to make 2 more points. One question is, should we even be using Child-Pugh as our way to risk stratify these patients? Child-Pugh is validated for prognosis in patients for cirrhosis, but there’s a lot of nuance there, and patients can be very different Child-Pugh Bs, as I mentioned. Perhaps the ALBI [albumin-bilirubin] score, which incorporates only the albumin-bilirubin, is a little more objective and something that we should definitely look at in future real-world studies.

The second point is that 1 of the components of Child-Pugh and ALBI is the bilirubin. I’d be curious to see what the rest of the panel thinks, but in our clinical practice, there’s a difference between these patients who have been decompensated for a long time with the bilirubin that’s been smoldering in the 3-to-4-mg/dL range, and they’re clearly decompensated vs a patient who has had good liver function and more acutely decompensates in the setting of tumor growth. Those would be patients in whom you’d consider overlooking that bilirubin—especially given the safety that Dr Gholam mentioned that we’ve seen with nivolumab in this population—and consider treating, with the rationale that as you shrink the tumor, perhaps you’ll see some improvement in the liver function.

Anthony B. El-Khoueiry, MD: That’s a great point. I’ll summarize by saying that the advice to our colleagues in practice is that if you have a Child-Pugh B patient, not all Child-Pugh B patients are equal. There’s a huge heterogeneity. Those patients would benefit significantly from a multidisciplinary discussion to assess whether there’s a potential benefit from treating their cancer. I’d like to emphasize that the major prospective data that have been reported for Child-Pugh B are with sorafenib from the GIDEON registry and a small subgroup in the initial phase 2 study. There are some safety data for sorafenib. There are safety data from nivolumab from CheckMate040, where there are Child-Pugh Bs in cohorts 7 and 8.

We know those agents can be used safely. They’re single-arm studies, but compared with historical control, there may be some benefit with survival. There are emerging data, very small data sets with atezolizumab-bevacizumab, with 1 showing decent safety and 1 from Korea showing some higher risk of bleeding. We need more data for the combination of atezolizumab-bevacizumab, and as Dr Gholam mentioned, more clarity on mitigation of bleeding risk in those patients. Lastly, the noninferiority of durvalumab to sorafenib recently brought back the option of single-agent PD-1, and that complements the data set with nivolumab in Child-Pugh B patients as well. Dr Li, before we move on from this subject, is there anything you’d like to add?

Daneng Li, MD: No, I absolutely agree with the group in the sense that there’s definitely heterogeneity in terms of our Child-Pugh B patients. As Dr Rich said, if it’s because of the underlying disease, then it’s potentially worthwhile to pursue aggressive treatment to try to control the cancer to hopefully improve that patient’s outcome.

Nicole Rich, MD: I’d add 1 more point. Some of the populations that have been unexplored in this area that were excluded from the trials are hepatitis B and C coinfected patients, and patients with HIV and hepatitis C coinfection. Those are some of the other areas of unmet need for further study, because we see a decent amount of coinfected patients in our clinical practice.

Anthony B. El-Khoueiry, MD: Sure.

Transcript Edited for Clarity