A 77-year-old woman presented to her primary care physician complaining of abdominal pain and fatigue. Ghassan K. Abou-Alfa, MD, discussed the case with a group of physicians.
During the Case-Based Roundtable event, Ghassan K. Abou-Alfa, MD of Memorial Sloan Kettering Cancer Center New York, NY, discussed the case of a 77-year-old patients with hepatocellular carcinoma.
ABOU-ALFA: What we’re bringing up here—and these are very important discussions—is if the biopsy would be implicative of the radiologic findings, especially the LR. [If] the mass was LR 4, would you change your mind?
Assuming a biopsy was obtained, do we do genetic testing, or do we require any further imaging?
Whoever answered “I don’t need the biopsy,” I presume they are basing that on the LR of the radiologist. Whoever answered that, for an LR 4, would you do the same, or would you think about [doing a biopsy] then?
ALI: I [would], because of the initial diagnosis, and it is helpful for the further management of the patient. And, if it is possible without any complication, it’s good to have a tissue diagnosis.
ABOU-ALFA: I 100% agree with you. I would favor a biopsy….The biopsy would be important because, for example, combined hepatocellular cholangiocarcinoma can occur, but [in] under 15% of patients. Could there be other reasons—for example, the value of the genetic testing? Yes, of course.
I’m curious, how many of you will get genetic testing in patients with hepatoceullular carcinoma [HCC]?
ALI: I will get it, because these days we are doing the next-generation sequencing [NGS] in almost all metastatic disease. That’s why I will also look for the PD-L1 at the same time.
MUNDIA: Is it helpful to do a liquid biopsy, like a Guardant360, in addition to tissue? Or would you just use tissue and send tissue for NGS?
ABOU-ALFA: Ironically, regarding the genetics, you can make an argument not to do genetic testing in patients with HCC because the potential for genetic alterations is minimal, even though I admit I do it. But remember, sometimes insurances might not cover it, and understandably so. Even if you are [getting the] tumor mutational burden to give a reason to use immunotherapy—remember, immunotherapy can be applicable for HCC anyway. I would not necessarily sweat too much about the genetic testing, but if you must do it and you have insurance covering it, by all means. Why not?
However, regarding liquid biopsy, no. The genetic testing you need is an add-on, and you need to have it first on the tissue itself before you get a liquid biopsy. [This is not necessarily applicable, but for example], let’s say there is an FGF19 alteration in a patient with HCC. You want to try to follow on the prognostication of it. I can understand why liquid biopsy would be done with Guardant360, especially if it correlates in that specific patient with outcome, even though there are still lots of questions about that.
MUNDIA: How is liquid biopsy compared with tissue for, say, an NTRK fusion in this disease? I think it’s unclear whether liquid is as good for fusions as it is the tissue.
ABOU-ALFA: To be fair, I cannot give you any data on that specifically, but I can give you data across the board. The data reported by Guardant on multiple occasions already— it appears they claim they have a high reproducibility of what’s in the tissue compared with the liquid biopsy.
But could it be that you’re comfortable [with] that said? Probably not. Maybe [it’s] good for diagnostic purposes, but if you [look at it from] a prognostic perspective, liquid biopsies need further work.
XU: I thought that we do not need a biopsy for some patients; based on imaging, we can make a diagnosis. I thought the situations we would usually use that would be hepatitis B positive, and this patient is not hepatitis B positive. But you can use 2 imaging modalities, and if you have this arterial enhancement and quick wash [out], and if that’s shown on 2 modalities, then you can diagnose HCC without imaging. Is that still correct?
ABOU-ALFA: All of what you said is correct, but I respectfully disagree to that [method] because the first imaging evaluations that were done to say [what] can be diagnostic in that regard were written…for screening purposes. Interestingly, at that time, exactly as you said, [there were] 2 imaging modalities. There was a statement…that patients should have cirrhosis, but [it was] never defined how you’re going to define cirrhosis. How we’re going to define cirrhosis is going to be a big subject that we’ll never finish talking about or discussing.
This was carried on as a prognostic perspective, then a lot of doctors started using it for diagnostic purposes, even though it’s not really the aim of it. Back in 2012 or 2013, I recall, [the author] said, “As much as I said this is for screening, it’s not for diagnosis, and because we need to learn [more about the disease] from the genetics, we need the biopsies.”
Then things evolved until the LR, and I spoke with the person who wrote the LR directly. I tried to learn from them. The answer that I got is “Yes, it’s a good language for me to talk to my other colleague radiologists.” The same way [oncologists] talk about stage IV, we understand each other, but we don’t necessarily go into these details or the confirmatory perspectives.
Same way, pathologists will say, “This is not due to carcinoma.” But at the end of the day, the pathologist is going to diagnose for patients with HCC. If anything, I strongly recommend we should not treat any patient with HCC without a biopsy.
XU: Wow, so that will change our discussion in our tumor board, because it’s dominating our tumor board—those 2 modalities—[so] we can make a diagnosis.
ABOU-ALFA: Who’s on that tumor board? Are you there all the time, or the medical oncologist, or your friends from radiology and the surgeons?
XU: No, some of our medical oncologists also think that way. But [it’s] important we send a message that that’s probably not right. But does that make a difference? We should get a diagnosis because it’s metastatic disease, but if a patient has resectable disease, then sometimes they don’t want to do a biopsy.
ABOU-ALFA: I’m glad that you go to the tumor board because I always notice—from what I hear—is that the oncologists are the least to go to the tumor board…Good for you that you can tell them, in our discussion that, for example, our colleague from Memorial Sloan Kettering [Cancer Center], they do biopsies. By all means, you can pass this information and support in that regard.
Then you bring the other point, which is, if it’s resectable, then yes, there’s no need for biopsy because there’s a resection going on.
The track of the needle is not a big deal to any further spread of the disease, even though I’m saying we don’t need biopsy for surgery, but not because of that, because, truly, we’re going to take it out anyway. With the tracking, the chance for you to spread the tumor from your biopsy is 0.000003%—this was reported in the literature already, so I would not worry that much.
ABOU-ALFA: A preferred regimen [in the National Comprehensive Cancer Network (NCCN) guidelines] is atezolizumab [Tecentriq] and bevacizumab [Avastin]. Other recommended regimens are sorafenib [Nexavar] and lenvatinib [Lenvima].1 Used in certain circumstances are nivolumab [Opdivo] and FOLFOX [folinic acid, fluorouracil, and oxaliplatin].
FOLFOX is already approved as a first-line therapy in Asia, specifically in China. There was an Oncologic Drugs Advisory Committee [ODAC] meeting that reviewed the data of nivolumab vs sorafenib. The FDA declined the nivolumab as first-line therapy.
WU: Why was nivolumab removed but pembrolizumab [Keytruda] was not? Because they would have similar data, either first line or second line.
ABOU-ALFA: The ODAC [voted] 5 to 4 in favor of the pembrolizumab. At the end of the day, the decision was made. We have to respect the different opinions. Different options count, but it was approved because it was in favor for the pembrolizumab. Of course, you can dissect the information and data any way you want, but the bottom line is it seems the separation for the pembrolizumab vs the placebo was probably more apparent and maybe why it favored the 5 who voted yes for the pembrolizumab and not for the nivolumab.
I want to go back to another question; at least 2 of you said atezolizumab/bevacizumab. I can understand why—because it is a NCCN preferred regimen. But I’m curious, was there no concern about this patient getting atezolizumab/bevacizumab?
ALI: The platelet count is adequate and there is no possibility of bleeding. Then you have 2 mechanisms of action. You’re using immune checkpoint inhibitor and vascular endothelial growth factor inhibitor. It is category A; that’s the reason I recommend this medication.
ABOU-ALFA: Fair enough. But interestingly, if you recall, this patient had variceal bleeds and Chron disease. In other words, you have 2 very important concerns: checkpoint inhibitors and bevacizumab, because the atezolizumab/bevacizumab study did not favor using bevacizumab in patients with a history of bleeding within 6 months. That’s why I would probably say it would be wise not to use it, because of that concern.
Why do any of you, if you do, [choose] not to use atezolizumab/bevacizumab as first line, other than if the patient has a risk factor of concern, like bleeding, varices, or any immune disease that can limit it? Is there any other reason why you would not do atezolizumab/ bevacizumab as first-line therapy?
ALI: If [the patient's] platelet count is low, then I will be careful.
ABOU-ALFA: I see. So if the platelets are low—it’s a very important point that platelets can be low. Does anybody have any argument as to why not, based on mechanism of action, interaction, or the data itself? “I don’t believe in it” or anything like that?
MO: Patient prefers oral medication instead of intravenous [therapy].
KHAN: [With] cardiovascular issues, recent pulmonary embolism, and things of that nature, you would not want to do bevacizumab.
KOBLENZ: I want to ask you about platelets, if I may. Most patients with cirrhosis have low platelet counts, which is defined as less than 150,000 mL. However, in my experience, you don’t get much bleeding unless there’s something else going on, unless the platelet count is less than 25,000 mL. Where do you draw the line regarding the risk for these agents? With regard to low platelets?
ABOU-ALFA: Admittedly, my tolerance level for the platelets for tyrosine kinase inhibitors can be in the 25,000 to 50,000 platelets. I agree with you, but I do agree with Dr Ali. If there is a concern about bleeding, you need platelets above 100,000 mL. This is understandable because of the potential for bleeding. Otherwise the platelets are not critical players, even though you must be careful. When the platelets are really low, it might be an implication that the cirrhosis is probably worse; ie, the Child-Pugh is probably B or C.
KOBLENZ: May I ask you about the dose reduction? If lenvatinib is so effective at 4 mg vs 8 mg or 12 mg, how much do you lose [in efficacy by dose reducing]?
ABOU-ALFA: Interestingly, at some point I remember there was an effort with the sorafenib [to ask]: “Why don’t you start at a low dose, and then increase the dose?”
This is wrong. Ironically, in the sorafenib days, I always started with a full dose, because the IC50 [half maximal inhibitory concentration] of the drug is very affected by that appropriate stoichiometric dose level. I’m saying that [lenvatinib is] effective at 4 mg every other day; I would say the same thing for sorafenib, which is also effective at 200 mg every third day.
But will you see the same efficacy—as robust— compared with the other [doses]? We will never have those studies done, but you need to start with the full dose, because this is where you can guarantee the effectiveness of the therapy that you need.
XU: Unresectable means it’s unresectable plus metastatic. What about those who are nonmetastatic and unresectable? Would you use systemic treatment first, or would you use…locoregional treatment?
ABOU-ALFA: When we decide which patient is amenable for embolization, we are automatically driven by “Yes, it’s local disease only in the liver. Let’s do what we have to do.” But if you look at it more carefully, you’ll notice that yes, there are a certain number of lesions in the liver; and yes, there is a certain size for those lesions in liver. Our colleagues in Japan came up with an interesting concept called the Up-to-7 criteria, where it’s easier for you to do.2
If you count how many lesions there is in the liver— let’s say 4—and you count the size of the largest lesion in the liver—let’s say 4 cm—so 4 plus 4 equals 8. They discovered that patients with more than Up-to-7 criteria will fare better on systemic therapy compared with local therapy.
They did a study and discovered that people would do way better on the systemic therapy than on the local therapy [if they had above 7]. [In] studies, survival with local therapy was always about 19 months. For patients with locally advanced disease, with more than Up-to-7 criteria, their median survival was 38 months on lenvatinib.3 I don’t think I need to say more. This is amazing.
XU: If those patients on lenvatinib had a 38-month survival—which is longer than the overall patient population for this nonresectable disease—does that mean that this cohort, which has not developed a metastatic disease yet, fare better if your systemic treatment is used earlier?
ABOU-ALFA: [Yes], if they are within the Up-to-7 criteria that we just referred to.
XU: But it also could be prognostic. Maybe they just didn’t have metastatic disease, they just progressed slower.
ABOU-ALFA: You’re absolutely right.
WU: What about patients who received transartierial chemoembolization?
ABOU-ALFA: [Survival was] 21 months.
WU: Would you offer lenvatinib for those patients? Would you offer atezolizumab/bevacizumab considering there are no contraindications?
ABOU-ALFA: Can you make an argument for doing the same approach for atezolizumab/bevacizumab? Yes, you can. I’m not going to say no, but the data we have are from the lenvatinib. This could be controversial for lenvatinib, even though we don’t have the direct evidence for that, but when you give lenvatinib, their liver performance remains stable.
Although they had some previous literature that if you give them the embolization, you worsen the liver functionality. Would it be the same for atezolizumab/bevacizumab? We don’t know, but that’s food for thought.
KHAN: What about combining locoregional therapies with systemic treatment?
ABOU-ALFA: Local plus systemic therapy all depends on what you are asking. If you are saying “I would like to use local therapy and throw in the systemic as an adjuvant therapy,” we did that in [multiple studies, and] they were all negative.
But the question can be turned the other way around: “What if I have a patient with metastatic disease, and I give them local therapy to the disease in the liver, and systemic therapy for the systemic disease?”
Interestingly, this was already tried in China, and they had great results that they presented at the American Society of Clinical Oncology annual meeting. Can I say this is what we should do? Not necessarily. Is it worth investigating further? By all means. I would say it’s an investigation perspective until we know more about it.
1. NCCN. Clinical Practice Guidelines in Oncology. Hepatobiliary cancers, version 4.2021. Accessed September 17, 2021. https://bit.ly/3Ao3MZW
2. Elshamy M, Aucejo F, Menon KV, Eghtesad B. Hepatocellular carcinoma beyond Milan criteria: Management and transplant selection criteria. World J Hepatol. 2016;8(21):874-880. doi:10.4254/wjh.v8.i21.874
3. Kudo M, Ueshima K, Chan SL, et al. Initial lenvatinib therapy with no prior TACE in patients with intermediate-stage hepatocellular carcinoma beyond up-to-seven criteria and Child-Pugh A liver function: A proof-of-concept study. J Clin Oncol. 2020;38(suppl 4):522. doi:10.1200/JCO.2020.38.4_suppl.522