Predictive Biomarkers for Checkpoint Inhibition Remain Unmet Need in HCC

Josep M. Llovet, MD

Despite recent developments in the treatment of patients with advanced hepatocellular carcinoma (HCC), there remains an unmet need for reliable biomarkers of response to immunotherapy, according to Josep M. Llovet, MD, who added that new regimens must also be developed to challenge the existing standards of care and further improve outcomes.

“In advanced HCC, [the challenge is to] identify responders to single-agent checkpoint inhibitors. Also, when we have a new treatment, we need to [think] 2 steps forward in terms of research,” Llovet said. “[We must] challenge this new treatment with a better combination or identify the ideal therapy for patients progressing to this combination, and whether this includes checkpoint inhibitors or not is up for debate.”

In an interview with OncLive during the 2021 International Liver Cancer Association Annual Conference, Llovet, who is the director of the Liver Cancer Program, and professor of medicine at the Mount Sinai School of Medicine, discussed the importance of identifying biomarkers of response to checkpoint inhibitors for patients with HCC, as well as future research directions.

OncLive: What is important to highlight from your presentations at the 2021 ILCA Annual Conference?

Llovet: [One presentation will address] the identification of biomarkers for predicting response to checkpoint inhibitors as single-agent [treatments] in HCC. Currently, the standard of care is atezolizumab [Tecentriq] plus bevacizumab [Avastin] in the frontline setting; other immunotherapies are approved by FDA, such as nivolumab [Opdivo]/ipilimumab [Yervoy] and pembrolizumab [Keytruda], both based on phase 2 data. In the phase 3 trials, nivolumab vs sorafenib [Nexavar] was negative, and pembrolizumab vs placebo was negative, so there is an unmet need for biomarkers to identify responders to checkpoint inhibitors.

In this study, we were able to dissect the profile of responders in a cohort of approximately 85 patients receiving single-agent checkpoint inhibitors for primary treatment of [patients with] HCC in the first- and second-line setting. We identified a gene signature of 11 genes that had an accuracy above 85% to predict responders and non-responders. Of course, this is just the first step; there is another group [of researchers] that also reported the signature that might be able to identify responders. Testing these signatures in prospective studies may be helpful to identify responders [to immunotherapy].

What are the immediate and long-term goals for this research?

Ideally, we would like to test [the signatures to confirm responders] in all cohorts who have been treated already with single-agent checkpoint inhibitors, [particularly] nivolumab or pembrolizumab. In the frontline [setting], approximately 20% of patients respond to single-agent treatment, so it will be very cost effective, in terms of treatment strategies, to identify [at least] most of the responders, and the others can be treated with a standard-of-care therapy. [Additionally, the median overall survival [OS] achieved with] the atezolizumab plus bevacizumab combination in the phase 3 IMbrave150 trial [NCT03434379] was approximately 19 months. [The median OS for patients with single-agent treatment is approximately 30 months], so if we are able to identify patients up front who might respond to single-agent [therapy], it will be very important step forward.

Pending further positive studies and results, what would be the clinical implications for practicing oncologists?

We have 3 unmet needs in terms of HCC treatments. The number 1 need at this point is to have any treatment as adjuvant therapy after resection or local ablation. Recurrence after resection or local ablation is a major problem in 50% to 70% of patients at 5 years, and there is no drug or treatment combination that has been able to prevent it. This is a major unmet need and there are at least 5 randomized control trials testing checkpoint inhibitors alone or in combination in the adjuvant setting. This will be a major breakthrough if any of these trials are positive.

In intermediate HCC, meaning patients with multinodular tumors in the liver only, chemoembolization [GM1] [has been the standard of care for 15 years]. There are several trials now combining chemoembolization with single-agent checkpoint inhibitors or in combination with TKIs or monoclonal antibodies. This will be also [pivotal for] improving survival for these patients.

What are the differences of managing toxicities of an immunotherapy with a VEGF TKI vs dual immunotherapy?

Based on the IMbrave150 trial, as well as reports on the combination of lenvatinib (Lenvima) plus pembrolizumab, and the combination of cabozantinib (Cabometyx) and checkpoint inhibitors, we can say that globally, checkpoint inhibitors plus bevacizumab have a slightly better profile in terms of grade 3/4 adverse effects [AEs], with up to 40% of patients reporting these complications. With checkpoint inhibitors plus TKIs, [the incidence these AEs] increases up to 60%. When deciding which treatment to use, the main decision-making is [often] based on outcomes, but AEs [should also be] in the decision-making process.

For patients who cannot receive atezolizumab plus bevacizumab in the frontline setting, could other single-agent TKIs, such as lenvatinib, be used?

Yes. In principle, the patients who have any contraindication for atezolizumab plus bevacizumab, either because they have autoimmune disease, or other issues that may prevent them using this combination, then first-line therapy should be either sorafenib or lenvatinib. Both drugs have been accepted as a standard-of-care up front, and they are accepted by all guidelines as primary treatment in those cases.

Regarding ongoing trials of triplet regimens in advanced HCC, how do you see this research playing out?

The triplet [regimens are] appealing, and from the pathogenic standpoint, it makes sense. There are 2 reasons [for this]. The first reason is that, so far, we have identified with the doublets that we are able to expand the target population that responds. [For example,] with single-agent checkpoint inhibitors, [the objective response rate is] 15% to 20%, but with combinations, we move that to 35% to 40%. Additionally, the disease control rate is very high with lenvatinib and pembrolizumab at up to 90%.

Secondly, the duration of response with combinations is approximately 12 months. However, what happens with the other types of combinations, such as nivolumab/ipilimumab? The response rate is smaller at around 25% to 30%, and the disease control rate also is about 50%, but the duration of response is longer.

Eventually, the ideal combination to capture both phenomena would be to have the wide range of a TKI, a PD-1 inhibitor, and then [enhancing] the depth of response by priming the PD-1 inhibitor with a CTLA-4 inhibitor beforehand. This is the strategy that has been used in the phase 3 HIMALAYA trial [NCT0329845], which is priming with tremelimumab before starting with durvalumab (Imfinzi). It is something that sounds appealing, and if the AE profile is sound and manageable, this can even further expand the response in patients.