Gholam Analyzes Treatment Outcomes for Advanced HCC in Child-Pugh B Population

Pierre Gholam, MD

Associate Professor, Department of Medicine

Case Western Reserve University School of Medicine

Cleveland, OH

PEERS & PERSPECTIVES IN ONCOLOGY: Could you summarize the design and goals of the phase 3 REFLECT trial (NCT01761266)?

GHOLAM: This was an open-label, phase 3, multicenter noninferiority trial that compared lenvatinib, which was prescribed in a weight-dosed fashion with 8 mg once a day for patients [weighing] less than 60 kg, and 12 mg in patients [weighing] 60 kg or more. The comparator arm was sorafenib [Nexavar]. These patients had to have at least 1 lesion you could see, and they primarily had to have Child-Pugh A status although there were a few protocol violations. They had ECOG performance status of 0 or 1. The primary end point was overall survival [OS] and secondary end points included progression-free survival [PFS], time to progression [TTP], and objective response rate [ORR], and there were some patient-reported outcomes that were also measured.

What were the efficacy and safety outcomes of REFLECT?

There was essentially noninferiority between lenvatinib and sorafenib, although numerically, the median OS for lenvatinib was 13.6 months and the median OS for sorafenib was 12.3 months with an HR of 0.92 [95% CI, 0.79-1.06].¹ The median PFS was 7.4 months for lenvatinib vs 3.7 months for sorafenib. The TTP was approximately 2.5 times that for lenvatinib vs sorafenib.

If you look at post-study systemic therapy, this is what patients got after they were censored in the trial, and 68% and 61% of the patients receiving lenvatinib and sorafenib, respectively, received nothing. After lenvatinib, 25% of patients received sorafenib; remember that was the only thing you could get [at the time]. Some patients [12%] after they received sorafenib received more sorafenib when they exited the study. There were other post-study medications; these probably included cytotoxic chemotherapy [such as] doxorubicin. [Findings from this study] led to [lenvatinib’s] FDA approval in the first line.

[In terms of grade 3 or higher adverse events (AEs)], hand-foot syndrome [occurred in] 3% of patients with lenvatinib vs 11% with sorafenib.¹ Diarrhea occurred in 4% [in both arms, whereas] hypertension was in 23% with lenvatinib vs 14% with sorafenib. Appetite loss occurred in 5% vs 1% and weight loss in 8% vs 3%, respectively.

I have prescribed tyrosine kinase inhibitors [TKIs] since well before these drugs became approved, to hundreds of patients. I’ve never had to interrupt a TKI of any kind, including this one, for diarrhea. There is always a very good strategy to manage diarrhea, primarily through the combination of an anti-motility agent [such as] diphenoxylate/atropine [Lomotil] or loperamide [Imodium] and a stool thickening agent [such as] non-natural fiber. Hand-foot syndrome is more of a problem in this study with sorafenib, and we know how to manage this because if you’ve treated patients with a TKI, that’s something that’s going to happen.

I try to take ownership of hypertension early typically by prescribing a calcium channel blocker for these patients myself as opposed to offshoring to a primary care physician. The biggest challenge is weight loss and appetite loss, which are cardinal symptoms of TKIs and there’s no magic bullet for that point. Increasingly we offer acupuncture as a treatment for cancer-related fatigue. There are some [findings from] randomized controlled trials that show it does work in some patients.

What data are there for frontline treatment of unresectable HCC with Child-Pugh B status?

[There was a comparison of] atezolizumab [Tecentriq] plus bevacizumab [Avastin] vs lenvatinib in patients with Child-Pugh B [status].² This was a study where there were 217 patients who received lenvatinib until atezolizumab plus bevacizumab’s approval. Then it was the physician’s choice to remain on lenvatinib or to switch to atezolizumab/bevacizumab. This was not a clinical trial; this was a real-world series of patients. In these 217 patients, 65 received atezolizumab plus bevacizumab and 70%, or 152, received lenvatinib, and we’re looking at those 2 groups.

In the patients who were treated with atezolizumab plus bevacizumab with data that were collected from different centers, the breakdown of etiologies and…patients who had previously undergone surgery were remarkably similar. Two-thirds had had Barcelona Clinic Liver Cancer stage C [BCLC-C] which means metastatic disease, and about a third had significant multifocal disease without metastasis [BCLC-B]. Most of these patients had ECOG performance status 0 or 1, and there was a relatively small percentage compared with [the population from] the IMbrave150 [NCT03434379] or REFLECT trials of patients who had portal vein thrombosis, and there were fewer patients who had extrahepatic disease, only about 28% [in the atezolizumab/bevacizumab group] and 38% [in the lenvatinib arm], although proportionally for both groups, they were very similar.

For that group of patients, the complete response rate in the lenvatinib arm was 4% and 3% for atezolizumab plus bevacizumab.² This is a small number of patients to begin with. The partial response rate was 31% and 50%, respectively. The ORR was 35% vs 18% and disease control rate was 65% and 55%, respectively. Again, this is not a randomized clinical trial; this is basically a case series of patients.

If one looks at PFS for the atezolizumab plus bevacizumab arm…for PFS…these are largely collated all the way to about 15 months. For OS there appears to be some separation of the curves in favor of lenvatinib at roughly 10 months [although] again the numbers are small.

Breaking [OS] down by BCLC-B and BCLC-C, for BCLC-B patients, it appears that you have somewhat more differential survival benefit, if you just look at the way the curves separate, than for BCLC-C, which means more advanced disease. Does that mean that you should be using lenvatinib in the intermediate stage, which is BCLC-B? Not necessarily, but what those data show in patients who have an earlier stage is that you may potentially have a group of patients who have a differential in survival benefit.

What other safety data are available for treating patients with Child-Pugh B status and worse?

There are some data in patients with Child-Pugh B status, based on a global registry called GIDEON looking at the real-world safety of sorafenib, and…this is by far the largest real-world data study ever done for HCC, of which most patients had Child-Pugh A [status] but approximately 700 had Child-Pugh B [status] and approximately 74 had Child-Pugh C [status].³ But the bottom line…is that it is remarkably safe to give sorafenib, even if they have Child-Pugh B7, B8, or B9 [status]. When you go into Child-Pugh C [status], and you look at survival, you are in very hazardous territory because patients with Child-Pugh C [status] have a 50% mortality rate just from having liver disease.

Beyond that, it is remarkable how well tolerated these are. For dose reductions, if you look in Child-Pugh A [status] it happened in 40%, in Child-Pugh B7 [status it was] in 31%, in B8 it was 30%, in B9 it was 25%, and in Child-Pugh C it was 26%. You’re not dose-reducing a lot more. Duration of treatment is going to be shorter as liver disease progresses because the patient is not going to live as long in general, and serious AEs of all grades were comparable among those different groups.

If you look at survival by Child-Pugh A score, [this would be the same] in a patient who does not have HCC, just liver disease. Patients with Child-Pugh A [status] have a greater survival than those with Child-Pugh B [status], which is about twice as much as Child-Pugh C [status]…if the patient did not even have HCC, so that’s not surprising. If you break it down by subgroups, there is a significant difference between B7 and B8. B7 has a longer survival compared with B8 and B9, which are with each other all the way out to 12 months, where there is some separation and then they catch each other [at 18 months].

How do patients respond to treatment if their Child-Pugh liver function worsens during treatment?

Because there are significant data gaps in the setting of Child-Pugh B disease, which many of us struggle with—because roughly 30% of my patients have Child- Pugh B [status] when I see them—people have been trying to dig up some data from existing randomized controlled trials to see whether we can glean some safety and efficacy data. This is one such effort from the REFLECT trial. These are patients who at 8 weeks had experienced progression from Child- Pugh A to Child-Pugh B [status] and remained on the study.⁴ There were approximately 60 of them. The Child- Pugh A subgroup, which is everybody else, was 413 [patients].

If you look at partial responses to lenvatinib, it was 28% for Child-Pugh B [status] vs 40% for Child-Pugh A, progressive disease was 20% vs 16%, and ORR was 28% and 43%, respectively [Figure⁴]. The median time to first objective response was 1.9 months and 1.9 months, respectively. Clearly there is some drop-off in efficacy if you have a sicker patient. The question is whether that drop-off in efficacy by itself is a reason not to offer this patient therapy.

Looking at PFS, the number of patients at risk over time is less favorable for Child- Pugh B [status] than Child-Pugh A. If you look at the patients receiving sorafenib, both groups perform less favorably than either lenvatinib group. If you look at the OS probability over time, the Child-Pugh A group performs better, and the OS probability is the lowest for patients who had sorafenib and Child-Pugh B [status].

The other dimension of this is safety. If you look at the Child-Pugh B subgroup, the mean daily dose intensity was 8.4 mg vs 9.5 mg, so it is lower, but I wonder how many of us have patients who receive 12 mg all the way. That’s not many patients in my practice, so 8.4 mg would be close to where I usually sit in terms of most patients. Duration of treatment was less [with Child-Pugh B status] because [patients with Child-Pugh A status] also tend to live longer; 3.2 months with Child-Pugh B [status] vs 6.9 months with Child-Pugh A. Serious treatment-emergent AEs were proportionately more in the Child-Pugh B group. The number of AEs [leading to] withdrawal were 15 for the Child- Pugh B group and 35 with Child-Pugh A, respectively. That would be 0.58 and 0.12 AEs per patient-year, respectively. [AEs led to] dose reduction in 46 vs 227 [1.77 vs 0.76 per patient-year], and dose interruption in 59 vs 275 [2.27 vs 0.92 per patient-year].

If you look at grade 3 or higher AEs in the Child-Pugh B group, there were more patients to develop ascites; that is why their [status progressed to] Child-Pugh B. But if you look at…diarrhea, fatigue, [etc], most of these are OK. The AEs that make a difference are all components of liver disease [such as] hepatic encephalopathy. More patients develop that in the Child- Pugh B group. Aside from that, [AEs] are more or less the same, including weight loss. This is the dataset that we have in that subset of patients that gives us a sense of their safety.

REFERENCES:

1. Kudo M, Finn RS, Qin S, et al. Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomised phase 3 non-inferiority trial. Lancet. 2018;391(10126):1163-1173. doi:10.1016/S0140-6736(18)30207-1

2. Rimini M, Persano M, Tada T, et al. Survival outcomes from atezolizumab plus bevacizumab versus lenvatinib in Child Pugh B unresectable hepatocellular carcinoma patients. J Cancer Res Clin Oncol. 2023;149(10):7565-7577. doi:10.1007/s00432-023-04678-2

3. Marrero JA, Kudo M, Venook AP, et al. Observational registry of sorafenib use in clinical practice across Child-Pugh subgroups: the GIDEON study. J Hepatol. 2016;65(6):1140-1147. doi:10.1016/j.jhep.2016.07.020

4. Huynh J, Cho MT, Kim EJH, Ren M, Ramji Z, Vogel A. Lenvatinib in patients with unresectable hepatocellular carcinoma who progressed to Child-Pugh B liver function. Ther Adv Med Oncol. 2022;14:17588359221116608. doi:10.1177/17588359221116608