Catherine Frenette, MD, explained in a Targeted Oncology live case-based peer perspectives presentation how she treats patients with hepatocellular carcinoma across the course of disease.
Catherine Frenette, MD
Catherine Frenette, MD, explained in aTargeted Oncologylive case-based peer perspectives presentation how she treats patients with hepatocellular carcinoma (HCC) across the course of disease. Frenette, a medical director of liver transplantation, Scripps Center for Organ Transplantation, Scripps Green Hospital and director of the Liver and Hepatocellular Cancer Program, Scripps MD Anderson Cancer Center, discussed the considerations she makes when deciding on treatment in the frontline based on the case study of a woman with HCC.
A healthy, 77-year-old Caucasian woman living in Utah with a history of alcohol use presented to her primary care physician complaining of abdominal pain and fatigue. She had an ECOG performance status (PS) of 1. Her medical history was remarkable for hypertension, which was well controlled on valsartan.
A CT scan of the chest, abdomen, and pelvis with triphasic liver evaluation revealed a 4.5-cm hepatic mass in the right lobe plus multiple small nodules in the lung. She was considered Child-Pugh class A. Her alpha-fetoprotein (AFP) level was 380 ng/mL and she weighed 65 kg.
Targeted Oncology: What are your initial concerns with this patient?
Frenette:I always worry, especially in older women, about labeling someone as having metastatic disease without knowing it is metastatic disease. For somebody like this, I would say her liver looks like there is hepatocellular carcinoma [HCC], so we do not need to biopsy that. But we need to biopsy the lung lesion if we do not have any old CTs we could dig up.
Targeted Oncology:Would you recommend a biopsy in this case?
Frenette:To prove metastatic disease, maybe just of the lung. Her AFP is 380 ng/mL; it is not 10,000, where I would be more comfortable giving her something systemic [without a biopsy].
I do not think it is wrong to biopsy the lung nodules. I look at every scan, and I talk to the radiologist and ask, “How suspicious are you?” Some of these you can look at and your radiologist will know it is metastatic disease, and then I do not biopsy it because even though a lot of the oncologists are sending biopsies for genetic testing of all these genes, we do not have enough data for me to have bought into that. They have looked at all these genes in HCC, and nothing they have found is predictive as of yet.
She is 77 years old. What if you biopsy it, and you drop her lung? Now she has a chest tube, she is in the hospital, and now she gets a pneumonia. She has an ECOG PS of 3, is taking a long time to recover, and you cannot give her systemic therapy.
Targeted Oncology: Would you ask for a PET scan for this patient?
A positive PET scan is helpful, but a negative PET scan is basically useless. Forty percent of HCCs are PET positive, and 60% are PET negative, so a PET is not going to help. It is a decent number, and it is predictive of vascular invasion and of poorly differentiated and more aggressive disease. There are data looking at patients who downstage to transplant or patients who have very high AFP levels, and you do a transarterial chemoembolization or ablation, and now their AFP is dropping, and you are trying to figure out whether you can do resection. If you do a PET, PET is predictive pretransplant of a posttransplant recurrence. Some transplant programs are now doing it for all downstaged patients to say if you are PET positive, we are not going to risk it.
Targeted Oncology: What are the first-line therapeutic options for this patient?
Let us assume that this 77-year-old woman is otherwise healthy and looks like she is 57 years old. Let’s also assume that you performed a biopsy, she did not have a complication, and she does have metastatic disease. She is Child-Pugh A. We are now talking about systemic treatment.
You have sorafenib [Nexavar], lenvatinib [Lenvima], or a clinical trial, if you have that available, as your systemic therapy options. I think the point of this is: How are you going to decide what you are going to use?
Targeted Oncology: How do you decide between these options?
If the patient is fit and has an ECOG PS of 0, I would give her a tyrosine kinase inhibitor [TKI]. I start low in those patients and ramp up the dose as opposed to just throwing the full dose at them because they are older. And for anybody who is a hint frail, I would start low and ramp up. For posttransplant patients, I start low and ramp up because we know that posttransplant, they do not tolerate the full dose as often.
There are differences in responses, but first I think about adverse effects [AEs] because these patients do have a deadly disease, and we want to maintain their quality of life. The main AE of lenvatinib is hypertension, and the main AEs of sorafenib [are] handfoot skin reaction [HFSR] and diarrhea, but that’s the balance. If I have a 77-year-old who is coming in, and he has hypertension, is on 3 antihypertensives, and is in my office with a blood pressure of 140/85 mmHg, I might shy away from lenvatinib and move toward sorafenib because I am not going to have as much trouble controlling the blood pressure.
Lenvatinib has a little more of a VEGF effect than sorafenib does. It hits VEGF1/2/3, and that is why you get more of the hypertension. However, there was a study at the [2019 Gastrointestinal Cancers Symposium whose results] showed that the people who had hypertension on lenvatinib had much better overall survival [OS] and stabilization of disease.1If I see hypertension, I think, “Awesome. This is goodit is working.”
And with sorafenib, [there is HFSR]. In San Diego, we have a lot of people who work outsidegardeners, construction workers, surfers, all sorts of people. They come in and already have the thickest calluses on their hands I have ever seen. If I shake somebody’s hand and it hurts me, I am worried about giving them sorafenib.
In the REFLECT trial, tolerability was a bit better with sorafenib, but in real life, it seems like people tolerate lenvatinib a bit better.2The lenvatinib AEs kick in faster, so I usually see them a week or 2 into treatment.
I am the one prescribing the TKIs because I have been doing it for 10 years. However, a lot of these patients are also receiving transplant, or they have some decompensation, their bilirubin is 2 mg/dL to 2.5 mg/dL, and the oncologist is worried they have encephalopathy. It is just easier because then I can manage the AEs and manage their cirrhosis at the same time.
Targeted Oncology: Before ramping up the dose, how long do you wait?
It depends on the agent. For lenvatinib, I usually give them about 1 to 2 weeks and I have them back in the office to see how they are doing. I get labs and ramp up the dose from there. If I am using second-line regorafenib [Stivarga], I do it in about a week, or 1 cycle, depending on how they are doing. Regorafenib is a 3-weeks-on, 1-week-off regimen, so if they are so-so, I might keep the dose where it is for the first 3 weeks. With sorafenib, the AEs take a little longer to kick in. It is more like 4 to 6 weeks, so if you try to ramp them up at 1 week, you are not going to be able to predict what their AEs are going to be a month later. I usually see them at 2 weeks, but I give them a full month before I ramp them up.
Lenvatinib is weight based. For patients who weigh less <60 kg, it is 8 mg; for >60 kg, it is 12 mg; and the dose reduction is by 4-mg doses. It is 12 mg to 8 mg to 4 mg to 4 mg every other day.
Lenvatinib 12 mg daily was initiated. She experienced modest weight loss and reported loss of appetite, for which she was referred for nutritional therapy. Imaging at 16 weeks showed a partial response to therapy.
Eight months after initiation of therapy, treatment was discontinued because of disease progression.
Targeted Oncology: Do you agree with the treatment, as well as the dosage, that she was given?
In a 77-year-old, I would have started at either 4 mg or 8 mg, depending on how biologically fit she looked and also how tolerant of AEs I thought she would be. Some people have AEs, and that is thatthey will never take the drug again. Other people will say, “I’m fine,” while their hands are falling off. You need to decide depending on the patient too.
Targeted Oncology: Why would you give this patient lenvatinib?
The REFLECT trial is the study that got lenvatinib approved for HCC specifically.2It is a noninferiority study. It was powered to show that lenvatinib was noninferior to sorafenib, which is smart because all the studies prior to it for first-line therapy versus sorafenib were powered for superiority, and they all failed. But that means because of the powering, you cannot look at the absolute numbers for OS and say, “It is better for OS,” because it does not meet that statistical significance even though it met numerical significance for OS.2
Another thing to realize about lenvatinib: It was not studied in patients who had massive liver involvement, defined as >50% liver occupation, or main portal vein invasion. Branch portal vein invasion was totally fine, but main vein invasion was an exclusion criterion. The patients were Child-Pugh A, but there were a few B7 patients. We just do not have safety data[that] is what that means. Once we get real-life data, we will.
I am comfortable prescribing systemic therapy in somebody with mildly decompensated liver disease because I know I can control it. I have Child-Pugh C patients who are working fulltime, and I have Child-Pugh A patients who will not tolerate a haircut, so I think about all of that, and I am not strict with the Child-Pugh A patients.
The primary endpoint was OS. Lenvatinib did beat the pants off sorafenib for progression-free survival [PFS]. PFS with lenvatinib was 7.4 months compared with sorafenib at 3.7 months [HR, 0.66; 95% CI, 0.57-0.77;P<.0001]; you get a doubling of PFS. Also, time to progression was about double, 3.7 months versus 8.9 months with lenvatinib [HR, 0.63; 95% CI, 0.53-0.73;P<.0001]. If you look at the median OS, lenvatinib was 13.6 months, and sorafenib was 12.3 months, and the hazard ratio was 0.92 [95% CI, 0.79-1.06], and it is not statistically significant, so this is noninferior. Now, if they had powered it higher for superiority, would lenvatinib have won?
We do not know because it was not powered for that. Looking at these numbers, you cannot, based on the REFLECT trial, say that lenvatinib beat sorafenib for OS. You can say it was better for PFS, better for response rate, and better for tolerability, but OS was equivalent.2
Targeted Oncology: What is the significance of tumor response to lenvatinib in the REFLECT trial?
The oncologists have been using RECIST 1.1 criteria forever, which basically looks strictly at size. It is not great for HCC, especially with liver-directed therapy. If you take somebody who has a 2-cm tumor, you ablate it, and you do a 1-cm margin around it, and now you have a 3-cm defect. That is a progression by RECIST 1.1 because the size is bigger. By mRECIST, it looks at only enhancementthat is all you count—so mRECIST is how we are talking about HCC, especially with the liver-directed therapies. Realize it still has not been completely validated for systemic treatment with HCC, so that is why they are always reporting both mRECIST as well as RECIST 1.1, but I always look more at the mRECIST than the RECIST 1.1.
[For the REFLECT trial], there is investigator mRECIST, independent mRECIST, and independent RECIST 1.1. With the investigator mRECIST, [patients on lenvatinib] had a response ratemeaning ≥20% tumor shrinkage—of 24.1%. By independent mRECIST, it was 40.6%, so that means that when they got somebody to have all the scans sent to 1 place that analyzed them all and did everything by their criteria that were carefully put out, this response rate was better than what we see with immunotherapy, which is about 16% to 18%.
Interestingly, what I am excited about is the bigger phase III studies of combination therapieslenvatinib plus immunotherapy. That will be interesting to get to.
We always think of sorafenib as a stability drug, but we do see some shrinkage and some decreased enhancement with sorafenib. I have seen some cases where I think, “Are you sure that is all?” But in independent review mRECIST, the response rate was still 12.4% with sorafenib.
In all the systemic treatments that we have right nowand we cannot compare trials head-to-head—we are seeing the best response rates with lenvatinib, and that is why we still want to stick with it for first-line therapy. Not that sorafenib is not a great option, because it is for certain patients. It is all about picking the right patient for the right drug and the right drug for the right patient.
They were able to keep people on treatment longer with lenvatinib, but they did have higher discontinuation for AEs if you look at the absolute percentage. That is because they kept people on the trial drug longer, so with sorafenib, they were on for 3.7 months, and with lenvatinib, they were on for 5.7 months. There was a 9% discontinuation rate versus 13%, but it was because of the duration that they were on treatment. The longer [patients] are on any of the TKIs, they get tired of it.
Targeted Oncology: Did many of the patients in the REFLECT trial require dose reductions?
I believe the total was 60% who had either dose reductions or dose interruptions, but 30% ended up staying on a reduced dose of lenvatinib. The data are the same with sorafenib, and cabozantinib [Cabometyx] is the same. In the SHARP trial, dose reductions occurred in 26% of patients on sorafenib, with dose interruptions in 44%.3