Minsig Choi, MD: What are some unmet needs in liver cancer? We have a lot of treatment options that have come up. Just like the questions that were raised previously—how do we sequence the prime therapies? Do we use sorafenib first because second-line options are for sorafenib failures? Do we use lenvatinib? Do we use checkpoint inhibitors in second-line setting or do you continue TKIs [tyrosine kinase inhibitors]?— that sequencing question becomes important. How we sequence currently effective liver cancer patients is 1 crucial question we have.
The second area of unmet need for my patient population is that most of the studies are done in Child-Pugh A patient populations. These are patients with good liver function, and I have seen in my practice about 30% to 40% of my patients actually are Child-Pugh B and C patients. Unfortunately, there are not many studies, except for nivolumab in Child-Pugh B7 patients. All other studies are done in Child-Pugh A5 and A6 patient populations.
Another unmet need is how we can manage those patients with poor liver functions and those Child-Pugh B patients. For Child-Pugh C we discuss hospice and palliative care, but in people with moderate liver dysfunction, how do we manage those patients? Some critical unmet needs are: Do we need to conduct clinical trials in patients with Child-Pugh or liver dysfunction? How do we sequence currently effective agents? How do we do clinical trials in other pathways and things? There is a lot of science and a lot of progress that still needs to be made in patients with liver cancer.
We are conducting a lot of clinical trials of adding immune checkpoint inhibitors in the earlier setting. For a patient after resection, we are looking at checkpoint inhibitors to see if those are effective in high-risk patient populations. We are also doing clinical trials looking at checkpoint inhibitors after a locoregional therapy, like after TACE [transarterial chemoembolization].
Those data will be emerging and trying to use systemic therapies earlier on in the adjuvant setting as well as in the locally advanced setting. But we also have to answer questions for patients with poor liver reserve. How do we manage those patients?
Masatoshi Kudo, MD, PhD: Most patients usually have poorly differentiated HCC. For those patients, we usually use the CT or Gd-EOB-DTPA [gadolinium–ethoxybenzyl-diethylenetriamine pentaacetic acid]–enhanced MRI scans to find differentiating factors. In that case, we performed a PET CT. PET CT is useful to diagnose poorly differentiated or invasive HCC, and the lenvatinib is very effective for poorly differentiated HCC. We use lenvatinib, which has good tolerability for poorly differentiated HCC, PET-positive HCC. Of course, lenvatinib is not perfect, so the very aggressive tumor is for those HCCs. We need another combination trial—a combination strategy like PD-L1 antibody plus anti-VEGF agent.
Peter Galle, MD: It’s diluting a lot along those lines. The problem in hepatocellular carcinoma patients is that the underlying cirrhosis is typically, over time, deteriorating. It is not just the tumor defining treatment options; it’s mostly liver function. We probably see only the sequence of 1 to the other line. This is what has been demonstrated by observational trials, that roughly a quarter only makes it from 1 line to the other.
Kate Kelley, MD: It was quite striking to see, from the CheckMate 459 trial, that around half of patients, after first line, went on to receive a second-line regimen. Likewise, from the KEYNOTE-240 trial in the second line, nearly half of patients went on to receive a third-line regimen. This is telling us that a subset of patients across lines of therapy are doing quite well and are able to try multiple approaches and to determine which is the most effective for their cancer. In some cases, this will allow us to perceive a third-line therapy, and I would add this is not an insubstantial number. This is a meaningful proportion of patients who will go on to receive third-line therapies.
That means that we need to better identify how to sequence therapies, particularly with the rapid changes in the frontline setting. We need to do our best using pooled analyses and retrospective analyses of patients who have received downstream therapies on clinical trials, as well as large real-world databases to the extent that they exist, to define the optimal approach in the later-line space. Ideally, we will have prospective trials to help guide choice of therapy. However, it will certainly take a long time and certainly won’t be feasible for every possible clinical question that we face. Therefore, it’s important to perform pooled and retrospective analyses, as well as look to the real-world data sets when available.
Masatoshi Kudo, MD, PhD: In the United States the checkpoint inhibitor monotherapies are approved: nivolumab and pembrolizumab. But in Japan, those 2 are not approved. We are waiting for atezolizumab-plus-bevacizumab approval. The IMbrave150 trial was positive and was published in the New England Journal of Medicine recently. It will be approved in Japan maybe this year.
The first-line option will be the atezolizumab-bevacizumab combination therapy. After that, we select the multikinase inhibitor, like sorafenib or lenvatinib, because there is a report that the PD-1 antibody binds to a CD8-positive T cell for more than 20 months. That means that even though progressive disease after immunotherapy is changing, switching to the multikinase-targeted agent like lenvatinib and sorafenib provided the very favorable situation that lenvatinib or sorafenib will kill the tumor cell and will change the tumor microenvironment, from tumor suppressive to tumor permissive. In that sense, for 20 months, there is that multikinase inhibitor, which provides the same situation with the combination therapy as immunotherapy plus TKI. Sequencing is similar after immunotherapy. In front line, maybe atezolizumab-bevacizumab therapy, then a first-line agent like lenvatinib, sorafenib, then ramucirumab or regorafenib or cabozantinib. Lenvatinib can also be used, as I said, third line and fourth line. It provides good response.
Kate Kelley, MD: My first hope is that we can identify a biomarker of response to immune checkpoint inhibition as monotherapy. There’s a reproducible number of patients who have deep and durable benefit from drugs like nivolumab, pembrolizumab, and even atezolizumab if we look at the phase 1b studies leading to the atezolizumab-bevacizumab regimen. Using immune checkpoint inhibitors as monotherapy, about 15% to 20% of patients can achieve a robust immune response. I hope there is a way to identify those patients, because it will help guide treatment sequencing, whether patients need an up-front combination or whether they could have this deep and durable benefit from a monotherapy, PD-1 or PD-L1 inhibitor, alone.
What that biomarker will be remains elusive right now. There is certainly a trend in many of these studies toward improved outcomes with higher PD-L1 expression on tumor, but not 1 study has shown a significant result, owing in part to small sample sizes and underpowering with respect to the biomarker question. This is because of small sample or perhaps an overall negative study end point.
There are other potential candidate biomarkers, such as immune infiltration or immune class gene signatures. Those are more complicated to scale as a biomarker across the larger population. We can look to data from other tumor types. It may be possible, at least based on the reproducible proportion of patients with a very clear-cut benefit. I hope there will be a monotherapy biomarker in the future.
Beyond a biomarker to define that population of patients likely to benefit from monotherapy, I see the treatment landscape evolving to include nuanced choice of combination regimens in front line, according to patient factors, tumor factors, and maybe a biomarker. Hopefully we will be able to develop wise approaches to choice of therapy in the second and third line using pooled and retrospective analyses to guide the sequence of therapy downstream.
Transcript Edited for Clarity