Novel Combos, Predictive Biomarkers Slated as Future of HCC Treatment

In Partnership With:

Partner | Cancer Centers | <b>USC Norris Comprehensive Cancer Center</b>

Anthony B. El-Khoueiry, MD, provides insight on the rapidly evolving treatment options in hepatocellular carcinoma.

Anthony B. El-Khoueiry, MD

With the armamentarium for advanced hepatocellular carcinoma (HCC) significantly expanding over the past few years, the next questions lie in the utility of biomarkers to determine which therapies are best for which patient, as well as the future role of combination therapies, said Anthony B. El-Khoueiry, MD.

“Our treatment has improved quite a bit because we have gone from having just one drug for about 10 years—sorafenib (Nexavar)—to now having a multitude of options,” said El-Khoueiry, an associate professor of clinical medicine at University of Southern California Keck School of Medicine, and the director of clinical translation at the Southern California Clinical and Translational Science Institute.

In addition to the plethora of available TKIs, more mature data could increase the prominence of PD-1 inhibitors, such as nivolumab (Opdivo) and pembrolizumab (Keytruda). The VEGFR2-inhibitor ramucirumab (Cyramza) is also now available for patients with alpha-fetoprotein (AFP) ≥400 ng/mL following treatment with sorafenib, added El-Khoueiry.

In an interview during the 2019 OncLive® State of the Science Summit™ on Gastrointestinal Malignancies, El-Khoueiry provided insight on the rapidly evolving treatment options in HCC.

OncLive®: How has treatment of patients for advanced HCC changed over the past few years?

El-Khoueiry: Let me define “advanced HCC” for a minute. The classic definition of advanced HCC is what we call Barcelona clinic liver cancer (BCLC)-stage C, which includes patients with portal venous invasion or metastases outside the liver. The other group that's included in advanced HCC systemic therapy trials is the patient population with multifocal disease limited to the liver, who have progressed on locoregional therapy or who are not candidates for locoregional therapy.

I'll take a moment to emphasize that this is a subgroup of patients with BCLC-stage B disease, meaning liver-limited disease, but they have large lesions diffused throughout the liver. These patients would not be served well with locoregional therapy so they go right to systemic therapy. This is the patient population that we typically include in trials for systemic therapy.

Starting in the frontline setting, we had a noninferiority study comparing lenvatinib (Lenvima) with sorafenib, and it was a positive study. It showed us that lenvatinib was noninferior to sorafenib as far as overall survival (OS). We are also awaiting the results of CheckMate-459, which is a study comparing nivolumab with sorafenib. These results are pending.

Then, we go to the second-line setting. The first positive study in this setting was the RESORCE trial comparing regorafenib (Stivarga) with placebo. This was a 2:1 randomization, and it was a positive study showing an improvement in OS compared with placebo. The secondary endpoints of progression-free survival (PFS) and overall response rate (ORR) were also superior to placebo. This population was unique; it comprised patients who tolerated sorafenib at 400 mg daily for the last 28 days, which was a very select group. Nonetheless, regorafenib is now an FDA-approved option following sorafenib.

Another agent that has been tested in the second- and third-line settings is cabozantinib (Cabometyx), another multikinase TKI. This drug targets VEGF, AXL, and MET, among others. This study did not select patients based on MET expression, [but it was] unique because about one-third of patients were being treated in the third-line setting. This was another positive study, and we again saw a median OS in the range of 10 to 11 months.

Ramucirumab is another agent that has been evaluated following treatment with sorafenib. This drug is an antibody against VEGF2. Ramucirumab was first evaluated in the REACH study, and this was a negative study. However, in a subgroup analysis, it appeared that in patients with an AFP ≥400 benefitted from ramucirumab. Therefore, the REACH-2 study [was started], which included patients with this characteristic, and all patients also needed prior exposure to sorafenib.

This was a positive study with ramucirumab showing an improvement in OS compared with placebo in this select population. A median OS of 8 months was a little bit less than what we were used to seeing, but cross-trial comparisons are a little bit unfair. There was a joint analysis of all the patients from REACH and REACH-2 with an AFP >400, and when these 2 trials are put together, we see approximately a 3-month improvement in OS with ramucirumab.

We also have immunotherapy agents. The PD-1 antibodies nivolumab and pembrolizumab both [received accelerated approvals [for use] following sorafenib. Nivolumab has an approval based on the CheckMate-040 trial; this was a phase II study with over 200 patients. The ORR with nivolumab was 20% when putting all of the cohorts together. When focusing on the post-sorafenib cohort, which is where the agent received accelerated approval, the ORR was close to 15%. What is important about these responses is that they were quite durable, with a median duration of response (DOR) reaching as long as 18 months. Looking at just that population, the median DOR was 10 months or longer.

Pembrolizumab was studied in a phase II trial with slightly more than 100 patients. We saw similar ORR of around 17% and the durability was comparable; toxicity profiles were also comparable. As I mentioned, we are waiting for results from CheckMate-459, which is the phase III trial evaluating nivolumab versus sorafenib.

How do you distinguish between lenvatinib and sorafenib?

REFLECT was a noninferiority study, so if we want to analyze the data from a purist's perspective, it's just another option; the OS was comparable. It's also interesting that the lenvatinib study excluded patients with >50% tumor burden in the liver, or with main portal vein invasion. When we look at secondary endpoints, the PFS and ORR were significantly higher with lenvatinib versus sorafenib; the ORR was in the low 20% range.

I am frequently asked, "When would you use lenvatinib?" It's hard to answer that question based on secondary endpoints. Nonetheless, this is quite a potent agent and the targets of lenvatinib are beyond just the VEGF receptor. Lenvatinib also targets FGFR, and this may offer unique mechanism of action. The higher ORR and higher PFS [seen with] lenvatinib [may make it] an option for symptomatic patients who need an immediate response or who may not make it to subsequent lines of therapy.

With regard to safety profile, these 2 TKIs actually had overlapping toxicities, except for 2 categories. There was a higher risk of hypertension with lenvatinib and a higher risk of hand-foot syndrome with sorafenib. If that matters to the specific patient, that could be taken into consideration as well.

What is your approach for choosing between immunotherapy agents and TKIs?

To elaborate a little more on the role of immunotherapy, the current data are based on phase I and II trials. The field is still evolving. Based on the data we have from the phase I/II studies, we could use nivolumab or pembrolizumab in the second-line setting, and they could also be used in the third-line setting because the eligibility did not have a limit on prior lines of therapy. Patients must have received sorafenib. It’s hard to know how to choose between a TKI and a PD-1 inhibitor in the second-line setting. There has not been a randomized phase III study comparing these agents in the second-line setting. A lot of this will be based on clinical judgement based on patient characteristics.

Let's look at the data and see how they may guide us. We know that with the PD-1 inhibitors, we give patients a chance for a long-lasting response. For those who may not be able to get to third- or fourth-line treatment, we may not want them to miss the opportunity to receive anti—PD-1 therapy. Maybe these patients are progressing rapidly or have high-risk disease, so we would give them immunotherapy in the second-line setting rather than a TKI. To be fair, the TKIs are FDA approved based on phase III data and level 1 evidence. We know that, with the sorafenib-regorafenib sequence in that select patient population, there are data that tell us if we look at the survival, this sequence had a median OS of 26 months. It's intriguing that sequencing these agents can give us survival beyond 2 years.

For those who may not make it beyond the second-line setting, we may want to preferentially use a PD-1 inhibitor. Some patients also may not tolerate TKIs, especially with the overlapping toxicities. Another angle to take here is the Child-Pugh class B patients, a population we always struggle with. Most of the data we have in clinical trials are from the Child-Pugh class A subgroup. We now have a cohort of 50 patients with Child-Pugh class B scores of 7 or 8 who were treated with nivolumab and the toxicity appeared comparable with the Child-Pugh class A population. The ORR was about 12%—so in the same range. This was a single-arm, small study, but it gives us some perspective.

What is some of the work being done with combinations in this space?

There are immunotherapy combinations of immuno-oncology (IO) agents plus other IO agents. The most obvious one in clinical trials right now is combining anti—PD-1/L1 with anti–CTLA-4. There is the phase III HIMALAYA trial combining durvalumab (Imfinzi) and tremelimumab versus durvalumab alone versus sorafenib in the frontline setting. We are awaiting data from a phase II cohort of nivolumab plus ipilimumab.

There is also rationale to combine PD-1 inhibitors with targeted agents against VEGF. For example, we can look at a pembrolizumab/lenvatinib combination, which has shown a promising early signal. We are also seeing nivolumab/sorafenib or nivolumab/lenvatinib in ongoing trials. We could go on and on. The question will be whether the toxicity is tolerable and whether the improvement in ORR will translate to better OS. The combination of atezolizumab (Tecentriq) with bevacizumab (Avastin) is also worth mentioning. This combination has data from slightly more than 100 patients, with a promising ORR in the 30% range. We will see again how this translates into survival.

Another important area of research is on biomarkers. Obviously, when we have multiple treatment options, it would benefit us to know how to select patients for the best therapy. There have been data emerging for the anti—PD-1 agents in terms of biomarkers. For example, in CheckMate-040, we showed that at least the PD-L1 expression was associated with higher ORR. This was not all or nothing, though—we still saw responders in the PD-L1–negative group. This is an evolving area, and none of these are definitive. I would not use any emerging biomarkers to select patients in my current practice.