During a Targeted Oncology™ Case-Based Roundtable™ event, Richard Finn, MD, discussed how treatment options for patients with hepatocellular carcinoma have advanced in recent years.
Targeted OncologyTM: How have treatment options advanced for patients with hepatocellular carcinoma (HCC) in recent years?
FINN: There has been a lot of progress in HCC. Historically speaking, if we look back prior to 2007, we were giving patients tamoxifen [Soltamox], doxorubicin, octreotide [Mycapssa], things that don’t work, because nothing had been shown to improve survival.
The first drug that improved survival was sorafenib [Nexavar], which was approved in 2007.1 That was based on the results of the SHARP study [NCT00105443], which was placebo-controlled because nothing was shown to improve survival. We saw that the hazard ratio [HR] was 0.69 for overall survival [OS], with a 31% decrease in the risk of death.2 Survival went from approximately 8 months to 11 months, and it did this not by inducing a high response rate but by slowing progression.
Then we had about a decade of negative studies, with no new drugs moving into FDA approval despite several phase 3 studies [of therapies] in the front line and in the second line. Then we had a bunch of approvals in the past few years. Regorafenib [Stivarga] was the first drug approved in liver cancer after sorafenib, and that was in the second-line setting compared with placebo.3 Nivolumab [Opdivo] got accelerated approval in 2017 based on single-arm data in the second line that showed durable responses in about 15% of patients.4 However, the phase 3 study of nivolumab vs sorafenib [CheckMate 459; NCT02576509] was negative, and therefore [nivolumab] was taken off the market.
Then the first frontline drug to be approved since sorafenib was lenvatinib [Lenvima], based on the results of the REFLECT study [NCT01761266], and that happened in 2018.5 Pembrolizumab [Keytruda] got accelerated approval in the second line.6 Pembrolizumab remains approved. Cabozantinib [Cabometyx] was approved in 2019 in the second line and third line vs placebo.7 Ramucirumab [Cyramza] was approved for the second line vs placebo.8 It’s the only drug associated with a biomarker, which is an AFP level of at least 400 ng/mL.
Nivolumab plus ipilimumab [Yervoy] was [granted] accelerated approval in 2020 in the second line based on the results of a single-arm study.9 Later in 2020, atezolizumab [Tecentriq] plus bevacizumab [Avastin] was approved.10 And [in late 2022], durvalumab [Imfinzi] plus tremelimumab-actl [Imjudo] was approved.11 We have a number of treatment options for our patients; 8 or 9 regimens are currently approved for HCC in the United States.
What were the study design and goals of the REFLECT trial of frontline lenvatinib?
This was a large, phase 3, open-label, global study that looked at lenvatinib vs sorafenib. The lenvatinib was dosed by weight : greater or less than 60 kg. The primary end point was OS, with numerous secondary end points. It is important to note that patients all had Child-Pugh scores of A. This study took patients with advanced HCC, Barcelona Clinic Liver Cancer stage C, and also patients who were intermediate [stage B], which means patients who had [transarterial] chemoembolization [TACE] but progressed and did not develop vascular invasion or extrahepatic spread.
They were still intermediate B, but they were refractory to TACE or they had liver-confined disease but had bulky tumors or infiltrative tumors that we know don’t do well with TACE. Importantly, this study had some exclusion criteria. Patients who had 50% [or more of] their liver involved or who had bile duct invasion were excluded, [and] an important fact is [that patients with] portal vein invasion at the main portal vein [were excluded].12 Liver cancers love blood vessels. First, they grow into the [nearby] blood vessels, which then lead to either the right or left portal vein, and then eventually [they grow] outside the liver into the main portal vein, which is Vp4. We know that patients who have main portal vein invasion do very poorly. This study was powered for noninferiority. That means lenvatinib could have been better than sorafenib, but [even] if it weren’t better, if it met this noninferiority threshold, it would [still] meet its end point.
What were the results of this study?
The OS on lenvatinib was about a month longer than on sorafenib [median OS, 13.6 months vs 12.3 months, respectively; HR, 0.92; 95% CI, 0.79-1.06]. The CI crosses 1, so it’s not superior. When they say “noninferiority end point,” that means the upper limit of the CI is less than 1.08, and here it’s 1.06. Therefore, it met the definition of noninferiority.12
The study met its secondary end points of improving progression-free survival [PFS], time to progression [TTP], and objective response rate [ORR]. The median PFS was 7.4 months on lenvatinib and 3.7 months on sorafenib, with an HR of 0.66. The HR for TTP was 0.63 [for ORR, P < .0001]. In calculating PFS and ORR, they used the modified Response Evaluation Criteria in Solid Tumors [mRECIST], which is similar to how we use Choi criteria for gastrointestinal stromal tumors. RECIST calculates ORR based on change in the sum of the longest diameters. You measure the longest diameter of target lesions, and then you see how [the sum of those diameters] changes over time.
If it goes higher than 20% from baseline, that’s progression. If it decreases by 30%, that is a [partial] response. [But] we know that tumors in the liver tend to be hypervascular, and sometimes drugs like lenvatinib and others can decrease the vascularity without decreasing the size of the tumor. So mRECIST takes into account only the size of the vascular component. All that being said, [the study] met its secondary end points.12
This benefit was very consistent across subgroups, [including those defined by] location, performance status, body weight, and AFP level. [The OS for] all of these tended to favor lenvatinib. Interestingly, the disease stage was prognostic, and for those patients with Barcelona stage B [disease], there was a clear survival benefit.12
This tells us that systemic treatment for patients with intermediate disease is important. We need to stay engaged with our patients. It’s fine and appropriate to send them to interventional radiology, but we need to see them back. Because if they just keep getting TACE, by the time they come to us they will be decompensated and [might not] benefit from medical treatment.
Now, the study [result] was “noninferior,” even though there were improvements in ORR and PFS. One reason was that there were more patients on the sorafenib arm who went on to receive poststudy systemic treatment. Patients who got lenvatinib were precluded from going on clinical trials because lenvatinib, at the time, was an investigational agent. In the lenvatinib arm, 68% of the patients didn’t get any poststudy medication, whereas [only] 61% on the sorafenib arm did [not receive poststudy medication]. A lot of the poststudy systemic treatment [consisted of] investigational agents that we now know are active in the second line, and that could have affected [the results of this study, such that it] did not meet superiority.12
Though many patients needed dose reductions, [the rate of] discontinuation for adverse events was fairly low, with no [significant] difference between the lenvatinib and sorafenib arms [13% vs 9%, respectively]. Patients were on lenvatinib for a longer period of time [median, 5.7 months vs 3.7 months for sorafenib], and [fewer of those patients received second-line therapy (33% vs 39%)]. And there was a delay in the deterioration of quality of life with respect to factors including role function, diarrhea, pain, and body [image] that favored lenvatinib vs sorafenib.12,13
What did the LEAP-002 (NCT03713593) study reveal about the combination of lenvatinib with pembrolizumab?
I was able to present this important study at the ESMO [European Society for Medical Oncology] Congress in 2022. The LEAP-002 trial came out of a single-arm study that showed that lenvatinib in combination with pembrolizumab had an ORR of 36% and a median OS of about 22 months.14 Lenvatinib is used in other diseases, and the combination of lenvatinib plus pembrolizumab has approval in kidney cancer15 and uterine cancer.16
The LEAP-002 study looked at the combination of lenvatinib and pembrolizumab vs lenvatinib plus placebo, so this was a double-blind, placebo-controlled study. Other phase 3 studies are typically open label because [in those studies] sorafenib is the control. This study was [conducted] in patients with Child-Pugh scores of A and, importantly, excluded patients who had main portal vein invasion, the Vp4 group.17
The study was well balanced. [Many] etiologies of liver cancer were represented. Patients all had a Child-Pugh score of A. Approximately 20% of patients in each arm were intermediate, whereas the rest were advanced. Most of the patients came from outside Asia.17
The primary end points were OS [and PFS]. The median OS with lenvatinib plus pembrolizumab was 21.2 months [HR, 0.84; 95% CI, 0.708-0.997; P = .0227].17 That’s a long survival for a frontline HCC study. The upper limit of the confidence interval was less than 1. However, because we had coprimary end points, and we had interim analyses that affected the statistics, in order to be superior, we had to have a P value of .0185. So we just missed the threshold for superiority. But, importantly, something that was totally unexpected was a median OS of 19.0 months with lenvatinib [plus placebo].
This was the first phase 3 study [of lenvatinib] since the REFLECT study, in which we had a median OS of 13.6 months, [and those studies had] similar inclusion criteria.12,17 And one would argue that if we used sorafenib as the control arm, which is consistently giving around 15 months [median OS] in recent phase 3 studies, this [result] would have been positive.
The OS [was assessed] in specific subgroups. Though many of these groups slightly favored lenvatinib, or [had an HR] just around 1, there were some groups that stood out. For example, patients with an AFP level greater than 440 ng/mL had an HR of 0.67. It’s important to note that high AFP is a negative prognostic factor. Similarly, patients who had macrovascular invasion or extrahepatic spread [favored lenvatinib plus pembrolizumab].
Again, these is a high-risk group of patients. Patients with extrahepatic spread had an HR of 0.78. Additionally, patients with hepatitis B [etiology] seemed to get a little better benefit [from the combination] compared with patients with hepatitis C or nonviral [etiology]. Had this study included patients with main portal vein invasion, making it a higher-risk group, that may have helped it be positive as well, because this regimen still had a high response rate.17
The RECIST[-assessed] ORR of the combination was 26.1%, and, if a patient responded, the response was very durable. Lenvatinib performed very similarly as it did in the REFLECT12 study, with an ORR of 17.5% [as assessed] by blinded independent review.17 [As assessed by] mRECIST, which can be tricky, the ORR of lenvatinib was 34.1%, but the combination had an ORR of 40.8%. It’s striking here that the combination performed very similarly as it did in the phase 1b study of 100 patients. The response rate in that study was 36%.14 So in a randomized phase 3 study, [the ORR] dropped, but the OS was very similar to what we projected. Again, the lenvatinib arm performed very well.17
Contributing to how the lenvatinib arm performed, 52.1% of patients on the lenvatinib arm went on to receive additional treatment.17 That’s higher than in the REFLECT study12 and higher than the lenvatinib plus pembrolizumab arm [in this study]. But, of that 52%, [22.8%] got additional treatment that was immunooncology [IO] based [Table17]. And that included IO plus VEGF therapy, presumably atezolizumab plus bevacizumab, which we know is a very active regimen, as well as single-agent IO.17