Atezolizumab/Bevacizumab Approval Leaves Second-Line Questions in HCC

Richard D. Kim, MD, discusses the evolving treatment landscape of HCC, the approval of atezolizumab/bevacizumab, and other pivotal data in HCC.

An influx of available agents, namely the combination of atezolizumab (Tecentriq) and bevacizumab (Avastin), has radically shifted the first- and second-line management of advanced hepatocellular carcinoma (HCC), said Richard D. Kim, MD, who added that further data are needed to optimize care in the second-line setting.

“We have to be more aggressive in managing these patients,” said Kim. “Oncologists should be involved in the treatment of [patients with] HCC much earlier than we have been in the past when we had nothing available. If a patient gets atezolizumab/bevacizumab in the first-line setting, the question becomes, ‘what do you do if they fail?’ We have to figure out how to sequence therapy. At the moment, atezolizumab/bevacizumab is considered [the] first-line [standard]. It may even move up earlier; however, we need to figure out what to do after.”

Atezolizumab/bevacizumab received regulatory approval from the FDA in May 2020 for the frontline treatment of patients with unresectable or metastatic HCC based on findings from the phase 3 IMbrave150 trial. In the trial, the combination led to a 42% reduction in the risk of death compared with sorafenib (Nexavar).

In an interview with OncLive during the 2020 Institutional Perspectives in Cancer webinar on Gastrointestinal Malignancies, Kim, a medical oncologist in the Department of Gastrointestinal Oncology at Moffitt Cancer Center, and an assistant professor of oncology at the University of South Florida College of Medicine, discussed the evolving treatment landscape of HCC, the approval of atezolizumab/bevacizumab, and other pivotal data in HCC.

OncLive: How has the treatment landscape of HCC evolved in recent years?

Kim: The treatment of advanced HCC has come a long way. A few years ago, the only drug that was FDA approved was sorafenib. Sorafenib was approved more than 10 years ago in 2007. That agent used to be the only standard of care treatment until recently. Now, we have multiple FDA-approved drugs available in the first- and second-line settings.

IMbrave150 was a phase 3 study that evaluated the combination of atezolizumab plus bevacizumab compared with sorafenib in the first-line setting. Those data were presented at the 2019 ESMO World Congress on Gastrointestinal Cancer, and updated data were presented this year. [The study was] also published in the American Journal of Medicine. Clearly, these data changed the landscape of how we manage [patients with] advanced HCC.

Could you discuss the IMbrave150 trial and how the combination of atezolizumab and bevacizumab has replaced sorafenib as the frontline standard of care?

In the past, sorafenib used to be the only available first-line TKI. Then came the REFLECT study, which was also presented a couple of years ago. That was a noninferiority study that showed that lenvatinib (Lenvima) was not inferior to sorafenib. However, the secondary end points showed a significantly longer progression-free survival (PFS) and a much higher response rate [with lenvatinib]. Based on that, lenvatinib has been taken up in the first-line setting.

Then, there was the phase 3 IMbrave150 study. This was a randomized trial in patients with unresectable HCC with Child-Pugh A disease. Basically, patients were randomized to the combination of the [PD-L1 inhibitor] atezolizumab plus the VEGF inhibitor bevacizumab versus sorafenib alone.

The key point here is that all patients who were on the study had to undergo endoscopy. The reason for that is that VEGF [inhibitors] can cause bleeding. A lot of patients with HCC have liver cirrhosis or hepatitis C, which puts them at risk for variceal bleeds. Therefore, patients [enrolled] in the study had to get [an endoscopy] done within 6 months and get their varices banned.

The study had a dual end point of PFS and overall survival (OS). During the first interim analysis, the investigators unblinded everybody because of the positive results. As I mentioned, the data clearly showed that there was an advantage in terms of response rate. The response rate with the combination was around 30%, and PFS was around 6 to 7 months.

Interestingly enough, the median OS for the combination has not been reached yet. Obviously, as more events occur, we will get [more data], but it seems like the median OS may be hovering around 20 months or longer. Clearly, this tells us that the combination is superior to sorafenib in patients who are fit to get the treatment.

In the past, we used TKIs in the frontline setting. Now we are using checkpoint inhibitors plus anti-VEGF agents in the first-line setting. The biggest unmet need is in the second-line setting. What do you do after patients [progress] on the combination? That is a question that has yet to be answered. By default, it is going to be a TKI. We don’t have any data on continuing on immunotherapy or a VEGF inhibitor beyond progression.

It is going to come down to using TKIs without much data. Either lenvatinib, regorafenib (Stivarga), cabozantinib (Cabometyx), or another TKI that we use in the second-line setting [will be used for patients who progress on the combination]. However, we need more trials in this setting to see what the right sequence of treatment should be.

This is one of the first trials to show superiority over sorafenib. Multiple attempts have been made to beat sorafenib, but none of them had. Lenvatinib was a noninferiority study, which met its end point; however, none of the combinations showed superiority to sorafenib.

What does the safety profile of the combination look like?

Overall, the combination was very well tolerated. Most of the adverse effects (AEs) were grade 1/2. Obviously, there were concerns about perforation or variceal bleeding. Some patients had variceal bleeding, but there were very few, comprising less than 5% of patients. Other than hypertension, the combination was very well tolerated.

Interestingly, a quality of life (QOL) [analysis] was presented, which clearly showed that despite the improvement in PFS, response rate, and OS, there was no compromise in QOL [with the combination]. Actually, one could make an argument that patients who were on atezolizumab/bevacizumab did a lot better in terms of QOL versus sorafenib. Sorafenib had other AEs, such as rash, fatigue, and hand-foot skin reaction that can be very cumbersome. Having said that, [the patients in the IMbrave150 trial] were Child-Pugh A and had good performance status.

As you mentioned, no data are available to guide second-line treatment for patients with HCC who progress on the combination of atezolizumab and bevacizumab. How are you approaching that decision in your practice?

By default, it is going to be one of the TKIs. The main reason for this is if you look at the indication for second-line therapy at the moment, a couple of agents are already approved, such as cabozantinib, regorafenib, nivolumab (Opdivo), ramucirumab (Cyramza), and pembrolizumab (Keytruda), as well as the combination of nivolumab and ipilimumab (Yervoy).

We understand that those studies were done in an era where sorafenib was considered the first-line treatment. Whether this can be applied when a patient progresses on atezolizumab/bevacizumab is unclear. One can argue that, at this moment, there are no data to show that a patient could continue on a PD-1 inhibitor once they fail a [PD-L1] inhibitor. Secondly, there are no data to show that a patient could continue ramucirumab once they have failed [on that agent]. There are data in colon cancer [that suggest] that we could use antiangiogenic agents beyond progression, but there are no data on that in HCC. Therefore, we are left with the TKIs that block the VEGF pathway, as well as other pathways.

The TKIs that are available in the second-line setting like regorafenib, cabozantinib, or lenvatinib may be [the agents] we use. However, moving to the future, there will likely be a combination study where we combine a TKI and continue on with immunotherapy. In other diseases, such as renal cell carcinoma, there are some data to show that if a patient fails a checkpoint inhibitor, we can continue with a different checkpoint inhibitor and add a TKI to resensitize those patients.

Could you discuss the CELESTIAL trial?

The CELESTIAL study was a randomized phase 3 study that evaluated cabozantinib, which is a TKI that blocks the VEGF pathway, as well as other pathways like AXL and MET. That agent is a bit different than other TKIs. The trial was done in the second- or third-line setting for patients who were intolerant of or progressed on sorafenib. The primary end point of the study was OS.

PFS and response rate were looked at as well. The study met its primary end point; there was an improvement in OS and PFS, and response rates were about 45%.

The unique thing about cabozantinib, at least in this study, is that it allows patients to fill 1 or 2 lines [of therapy]. A few patients in this study received immunotherapy, which is unique because most other studies didn’t include patients who received prior immunotherapy. The second thing that is different from other trials is that this trial allowed all-comers, irrespective of whether the patient’s tumor was large or whether the patient had portal vein thrombosis; the trial [had very few exclusion criteria].

[Investigators] did a post hoc analysis of patients who failed 1 line of treatment. As expected, the data show that with secondary therapy, patients [who receive 2 or fewer lines of therapy do better than] patients who receive more than 2 lines of therapy. Therefore, the OS and PFS in a pure second-line setting. However, the trial did include patients who failed 1 or 2 lines of therapy. It is a little bit different than other studies, but clearly, cabozantinib has activity in this setting. There are some toxicity concerns with [cabozantinib]; however, they are very similar to [the toxicities we see with] TKIs, such as hand-foot skin reactions, gastrointestinal toxicities, and hypertension.

What data have we seen from the REACH-2 trial, and what is the role of ramucirumab in HCC?

REACH-2 evaluated ramucirumab in the second-line setting for patients who failed sorafenib with an alpha-fetoprotein (AFP) level greater than 400. The trial came out after the REACH-1 study, which [evaluated ramucirumab] in all-comers in the same setting. REACH-1 was a negative study; there was no difference in OS. However, [in a post-hoc analysis], the investigators found that patients who had an AFP level greater than 400 benefitted from ramucirumab. Therefore, based on that hypothesis-generating post-hoc analysis, they did the REACH-2 study. REACH-2 was similar to REACH-1, but only patients with AFP greater than 400 were included. [Placebo served as the control arm] because nothing was considered standard in the second-line setting.

The results showed that patients who had an AFP level greater than 400 and had failed sorafenib had an improvement in PFS and OS with ramucirumab. The unique thing about ramucirumab as a single agent is that it is not a TKI and it blocks really only 1 pathway. Even though [the nearly] 2-month OS difference may be considered modest, the agent is very well tolerated. It is a single agent with very few grade 3/4 AEs. Most toxicities are grade 1/2 and are mainly hypertension.

Therefore, for patients with an AFP level greater than 400 who progress after sorafenib, ramucirumab could be considered an option.

What is your take-home message for your colleagues?

The treatment of patients with advanced HCC has evolved dramatically in the past. We have immunotherapy, targeted therapy, and combination therapy. In the past, where we only had sorafenib, we would treat patients with local therapy alone and give sorafenib at the end. Now, with the treatment [options] that are available, it is [important] that we choose the appropriate patient to send to oncologists earlier on, so we can treat them accordingly.


FDA approves Genentech’s Tecentriq in combination with Avastin for people with the most common form of liver cancer. News release. Genentech. May 29, 2020. Accessed August 18, 2020.