The optimal integration of new drugs into the treatment paradigm of HCC requires several key developments, including greater research into biomarkers that would help inform therapy choices, earlier treatment for patients diagnosed with HCC, and clinical trials that test combination strategies.
Ghassan Abou-Alfa, MD
The rapid expansion of therapeutic options for patients with hepatocellular carcinoma (HCC) during the past 2 years is inspiring, but optimal integration of the new drugs into the treatment paradigm requires several key developments, according to a panel of experts. These include greater research into biomarkers that would help inform therapy choices, earlier treatment for patients diagnosed with HCC, and clinical trials that test combination strategies.
Those were among the conclusions reached by leading HCC investigators who participated in a recent OncLive Peer Exchange® program. “We are experiencing a sea change in terms of systemic therapy for advanced HCC,” said Ghassan K. Abou-Alfa, MD, who served as moderator.
Sorafenib (Nexavar) was the only choice of systemic medical therapy for patients with advanced HCC until April 2017, when the FDA approved regorafenib (Stivarga) as a second-line therapy after sorafenib.1 Since then, the FDA has approved lenvatinib (Lenvima) as a first-line treatment for unresectable HCC2 and 3 drugs for patients who have previously received sorafenib therapy: nivolumab (Opdivo)3, pembrolizumab (Keytruda)4, and cabozantinib (Cabometyx)5. Additionally, the National Comprehensive Cancer Network recommends ramucirumab (Cyramza) for patients with α-fetoprotein (AFP) ≥400 ng/mL.6
Abou-Alfa, along with panelists Richard S. Finn, MD; R. Kate Kelley, MD; and Andrew X. Zhu, MD, PhD, reviewed noteworthy abstracts presented at the 2019 Gastrointestinal (GI) Cancers Symposium in January. They also discussed data published in 2018 that led the FDA to approve pembrolizumab in November 2018 and cabozantinib in January 2019 for previously treated HCC.4,5
Emerging Approaches With Immune Checkpoint Blockade
The PD-1 inhibitors nivolumab and pembrolizumab are currently the only FDA-approved immune checkpoint inhibitors (ICIs) for HCC. The FDA granted accelerated approval to nivolumab for advanced HCC previously treated with sorafenib in September 20173 and granted the same conditional approval to pembrolizumab a little more than 1 year later.4 The panelists noted that the FDA based the approvals on data from phase II trials that observed response rates of approximately 15% and an encouraging duration of response.
The panelists mentioned ongoing clinical trials evaluating the FDA-approved PD-1 inhibitors alone or in combination with other HCC agents in various settings. They also discussed novel ICIs being investigated for HCC, including the PD-L1 inhibitors atezolizumab (Tecentriq) and durvalumab (Imfinzi) and the CTLA-4 inhibitors ipilimumab (Yervoy) and tremelimumab. Although trials have reported promising response rates and duration of disease control in the second-line setting, “no data yet [show] that using these drugs improves survival in any setting,” Finn noted.
Approval of pembrolizumab for advanced HCC in patients previously treated with sorafenib was based on results from KEYNOTE-224, a phase II single-arm trial for which Zhu was lead investigator.4,7
“In the whole cohort of 104 patients, we observed a very solid 17% response rate,” he said. Responses tended to be durable, allowing some patients to stay on the trial for a prolonged period, he said. Investigators reported that 16% of patients who received at least 1 dose of pembrolizumab were still receiving treatment at the data cutoff. Almost one-quarter of patients experienced a grade 3 treatment-related adverse event (AE) during the study, with liver enzyme increases and fatigue being the most common.7
Shortly after the Peer Exchange panel discussion, Merck announced that the phase III confirmatory trial of pembrolizumab— KEYNOTE-240—failed to meet its coprimary endpoints of statistically significant improvement in overall survival (OS) and progression-free survival (PFS) compared with placebo.8 KEYNOTE-240 included 413 patients with advanced HCC previously treated with systemic therapy.8 The data have yet to be peer reviewed, and the ramifications of the negative findings are unclear.
Checkpoint Inhibitor Combinations
Clinical trial findings demonstrated promising activity with anti—CTLA-4 and anti–PD-1/PD-L1 combinations in other tumor types, which led to their investigation for HCC, according to Kelley. “We’ll look eagerly to those results to see if we can also harness the combined checkpoint inhibition for higher response rates…as has been seen in other tumor types,” she said.
At the 2019 GI Cancers Symposium, data were presented from a phase II study that evaluated a durvalumab/tremelimumab combination in a subset of 10 patients with advanced HCC. The partial response rate was 20%, and the disease control rate was 60%. Median PFS and OS were approximately 8 and 16 months, respectively.9 Abou-Alfa said the authors found the combination was well tolerated, and it is now being tested in the phase III HIMALAYA trial.
Although prior investigations have shown that adjuvant use of targeted agents does not improve outcomes in HCC, Zhu said the activity of PD-1 inhibitors in patients with advanced disease has prompted researchers to consider whether ICIs might be more efficacious if introduced earlier. A phase II study presented at the 2019 symposium assessed what the authors described as perioperative immunotherapy.10 In the neoadjuvant phase, patients with resectable HCC were randomly assigned to receive nivolumab alone or combined with ipilimumab over 6 weeks. They underwent resection within 4 weeks of the last immunotherapy cycle and continued to receive the same adjuvant regimen for up to 2 years.10
The interim analysis included data for 8 patients (5 treated with nivolumab monotherapy and 3 with nivolumab plus ipilimumab). “The point that really impressed [the investigators] the most was that there was…a complete response [rate] of 37.5% with regard to the pretreated patients who got to the resection,” Abou-Alfa said. Complete responses occurred in both the monotherapy (2 of 5) and combination arms (1 of 3).10 AEs did not preclude anyone from undergoing surgery, and no one experienced a grade 4 or higher AE. Three of 8 patients grade 3 AEs (2 enzyme increases, 1 colitis).10
Zhu praised the exploratory study for being well conducted and using an “intriguing and innovative approach.” He said it offered investigators a great opportunity to evaluate the effects of checkpoint blockade on the tumor microenvironment. Despite the encouraging results, Zhu said it was premature to use ICIs in the neoadjuvant or adjuvant phase for patients with HCC. “We have to see how the data are evolving in the adjuvant setting,” he said.
Other studies are assessing concurrent use of an ICI and a tyrosine kinase inhibitor (TKI). Kelley said preclinical data suggest “TKI therapy may modulate the immune microenvironment around the tumor,” that making it more responsive to checkpoint inhibition. At the European Society for Medical Oncology (ESMO) 2018 Congress, findings were reported from a phase IB study that combined atezolizumab with the antiangiogenic agent bevacizumab (Avastin) for advanced unresectable or metastatic HCC. The overall response rate (ORR) for the 68 evaluable patients was 34%.11 Although this was substantially lower than the 62% ORR reported in a previous interim analysis based on 21 evaluable patients,12 Zhu called it very respectable. He said the strong activity and manageable safety profile have led to further investigation of the atezolizumab/bevacizumab regimen in the phase III IMbrave150 trial (NCT03434379).
“Thirty percent in liver cancer—that’s a great response rate,” Finn said. He explained that the responses were confirmed by independent review using conventional RECIST criteria: “So, we’re really causing tumors to shrink.”
Finn agreed that the combination was highly active in liver cancer but said the real question that remains unanswered is, “Will this activity be better than sorafenib?”
New and Emerging Tyrosine Kinase Inhibitors
The FDA approved lenvatinib (Lenvima) in August 2018 as a first-line agent for unresectable HCC based on results of the REFLECT trial, a phase III noninferiority study that compared lenvatinib with sorafenib.2 OS was 13.6 months with lenvatinib versus 12.3 months with sorafenib, meeting the study’s primary endpoint of noninferiority.13 Finn, a REFLECT investigator, said that “on all of [the secondary endpoints], there was a benefit to lenvatinib—better control of tumor progression radiographically as measured by time to progression, significant increase in PFS, and a significant increase in response rate.” Lenvatinib was also associated with significantly greater tumor shrinkage.14
Finn said a post hoc analysis of OS data from REFLECT presented at the 2019 symposium showed that “if you have a response to either sorafenib or lenvatinib…your survival is better than if not.”15 Median OS was 22.4 months for people who responded to either treatment versus 11.4 months for nonresponders.15 The analysis also identified treatment with lenvatinib versus sorafenib as an independent predictor of survival.15 “The high [AFP] patients seem to get a little more benefit from lenvatinib,” Finn said.14
Kelley said her group started using lenvatinib shortly after it received FDA approval. “The main toxicities that we’ve encountered have been hypertension and some edema,” she said. The panel discussed differences in the safety profiles of sorafenib and lenvatinib, including an increased risk of ascites and hypertension with lenvatinib versus a greater risk of hand-foot skin reactions and diarrhea with sorafenib. Because all TKIs approved for HCC cause hypertension, it is important for patients “to have really strong blood pressure control prior to starting the drug,” Kelley said.
The FDA approved cabozantinib for previously treated advanced HCC based on outcomes in the phase III CELESTIAL trial, for which
Kelley and Abou-Alfa were lead investigators.5,16 Abou-Alfa marveled that cabozantinib received approval barely 1 year after data were first presented at the 2018 GI Cancers Symposium. “It’s a great other choice of therapy for our patients,” he said. Kelley said that although cabozantinib is like TKIs sorafenib and regorafenib in targeting VEGFR isoforms, it also targets MET and AXL.16 The panel agreed that the significance of targeting MET is unclear.
In the CELESTIAL trial, 707 patients were randomly assigned 2:1 to cabozantinib or placebo.16 Cabozantinib significantly prolonged median OS compared with placebo (10 vs 8 months; P = .005), as well as median PFS (5 vs 2 months; P <.001), and ORR (4% vs <1%, respectively; P <.009.16 Finn said he saw little difference in outcomes between cabozantinib in the CELESTIAL trial and regorafenib in its pivotal trials. “The data look very similar as far as tumor control, PFS, and adverse effect profile,” he said.
At the 2019 symposium, the investigators presented a post hoc analysis of quality-adjusted life years (QALYs) in CELESTIAL.17 They concluded that although cabozantinib was initially associated with a small reduction in QALYs compared with placebo, as the trial progressed, the health utility of cabozantinib progressively increased and was significantly better than that of placebo at the end of the study.17
The recommended cabozantinib dose is 60 mg once daily.5 Zhu said he has prescribed cabozantinib for select patients, some of whom developed what he believed were dose-related toxicities. He said the optimal dose for such patients was unclear but expressed the view that by “lowering the dose to 40 mg, these patients will tolerate it better and also get longer drug exposure.”
Ramucirumab, a novel VEGFR2 inhibitor that the FDA has approved to treat other GI cancers, is being investigated for HCC. REACH-2 is a phase III study comparing ramucirumab with placebo as second-line therapy for patients with advanced HCC and high AFP expression (≥400 ng/mL).18,19 Zhu, an investigator for REACH-2, said that patients with high AFP “have very poor prognosis, and they definitely fare very poorly.” REACH-2 data presented at the 2018 symposium associated ramucirumab with significant improvement in OS compared with placebo.19
At the 2019 symposium, investigators presented data from a subgroup analysis that showed that significantly more patients in the ramucirumab arm than in the placebo arm had an AFP response (≥20% decrease in AFP level).18 Patients with an AFP response had significantly longer OS than patients without an AFP response. Ramucirumab was associated with longer time to AFP progression and longer time to radiographic progression.18 Finn, also a REACH-2 investigator, said the findings show that AFP is prognostic and “predictive of response to ramucirumab.”
The panel agreed that the rapid transition from 1 available treatment option to 7 was inspiring but that many questions remained unanswered. “It’s time to step back and to really try to refine our understanding of the biology of this incredibly heterogeneous patient population,” Kelley said. She called for greater effort to identify or refine biomarkers that could aid in sequencing treatment. “Whatever we can do to really understand which patients respond to which strategies,” she concluded.
Finn said, “The only way we’re going to get patients to benefit from these drugs [is] if they’re physically in condition to be treated.” That means getting involved with patients earlier in the course of their disease and “probably not do[ing] that extra transarterial chemoembolization if it’s not indicated,” he said.
“We should not be content with having 7 drugs,” Zhu said. He mentioned the need to look for more drivers of HCC that could lead to new therapies. He also called for moving away from “the empiric combination strategy to some more rationally designed combinations” to improve the likelihood of success.