Treatment Strategies for Childs-Pugh B Disease

Richard S. Finn, MD: Another unmet need with all this systemic treatment is the idea that all these studies are of patients with Child-Pugh A disease. In the real world we see patients who are Child-Pugh B, and the issue is in part safety and efficacy. Laura, can you break that down for us, systemic treatment in less compensated patients?

Laure M. Kulik, MD: Not every patient who is Child-Pugh B is the same. Why are they Child-Pugh B? Is it ascites? Is it their bilirubin elevated or their albumin low? These characteristics may make that patient look physically different to you. There have been some data looking at some of the systemic agents in patients who are Child-Pugh B, and all were from small studies. The largest studies we have are with sorafenib and the GIDEON trial, in which patients who were Child-Pugh B had similar responses but their overall survival was poor likely because of the progression of their underlying cirrhosis. In patients who were Child-Pugh C, it was only 2.1 months median survival, so probably not worth doing.

For lenvatinib, at least in studies from the West, there has been a cautionary note that patients who are Child-Pugh B at initiation may experience higher rates of treatment discontinuation due to AEs [adverse events] and also increased the risk of developing jaundice.

The companies have tried to use the registration trials to look at the migration of patients to other classifications. For example, a patient started as Child-Pugh A. At the first point in the trial, which was week 8, how many of those patients became Child-Pugh B? What happened to those patients? With the lenvatinib, it’s a little curious that the patients who became Child-Pugh B had a response, whereas the patients treated with sorafenib who became Child-Pugh B didn’t have a response. Did they become Child-Pugh B as a result their tumor progression and the drugs not working as opposed to the patients who had lenvatinib and the drug was working? Are they becoming Child-Pugh Bs as an AE of the drug? These are all things that we have to sort out. Do we need to use lower doses of the drugs because it is hanging around? Will that mitigate some of the potential effects on liver dysfunction? These are all the things that need to be answered in this unmet need.

Richard S. Finn, MD: Yes, and I find with drugs that are inducing higher responses, to your point, some patients have a very large tumor, which is giving them this Child-Pugh B physiology. You look at their film, and their liver is not necessarily very shrunken and scarred and still doesn’t appear as cirrhotic on imaging. Sometimes you induce a response and their liver function improves. It’s an evolving concept for us given that we have drugs now that have durable and high response rates.

Josep, before we leave this section, alpha fetoprotein [AFP] has been around a long time. Does that weigh into your treatment decisions?

Josep M. LLovet, MD: AFP greater than 400 ng/mL has been established as a prognostic factor and a predictor for ramucirumab response. Eventually the indirect question is how do we use AFP in the strategy for choosing treatment? We were talking about frontline atezolizumab-bevacizumab, potentially first- or second-line sorafenib or lenvatinib, and we are talking about cabozantinib, regorafenib, and ramucirumab.

I want to point out 1 concept that this follows. This builds on the mechanisms of action [MOAs] of these drugs because I have been working on that, and I’m a bit skeptical about the knowledge that we really have. I’m not sure that because patients have been exposed to bevacizumab, the VEGFA ligand monoclonal antibody, that they cannot be exposed to a VEGF receptor 2 monoclonal antibody inhibitor, which has a more potent effect, certainly.

I want to drop here the data that all of you are familiar with. When we were testing regorafenib in the second-line setting, we were testing regorafenib in patients exposed to sorafenib. Regorafenib is a more potent sorafenib with a nanomolar IC50 [50% inhibitory concentration] but also has type 2 as an additional target.

The feeling was: Why are we testing regorafenib in patients who progressed to sorafenib? The MOA is similar and the hazard ratio was 0.63. Talking now about before ramucirumab, and this goes also for all the VEGF inhibitors in the second-line setting—for instance, cabozantinib and regorafenib work well in patients with high AFP—I would say that you are looking for a very good safety profile. You can select 1 timepoint, and if the patient is in good shape, you can select ramucirumab for patients with high AFP. It’s getting difficult, but it’s possible.

That builds on the concept of using checkpoint inhibitors in the second- or third-line settings. In the second line you may have changed the microenvironment or even in the third-line after treating with TKIs [tyrosine kinase inhibitors]. We’re sending a poster to ESO [European School of Oncology] where we didn’t know what happened to the TKIs in the microenvironment. Some tumors may switch from cold to hot later, but at the beginning they are not responding and later on they are responding, or they are not responding to atezolizumab-bevacizumab but respond later to lenvatinib-pembrolizumab because the MOAs are different and you are seeking a different population of tumors. I will be a little skeptical to completely rely on the MOA. It’s important, but we need to be cautious and gather more data.

Richard S. Finn, MD: Yes, we need to keep an open mind. We all have preconceived notions on how things work on our expectations based on MOA, but until there are data, we need to keep an open mind. There is an ongoing single-arm study with ramucirumab to see its activity and safety after other frontline therapies other than sorafenib and some of that will probably include atezolizumab-bevacizumab.

Transcript Edited for Clarity