Introduction

DR KELLEY: So good evening. I’d like to welcome you all to Cases from the Community: Investigators Discuss Available Research Guide in the Care of Patients with Hepatobiliary Cancers. And this is part 3 of a 3-part CME symposium series held in conjunction with the 2023 ASCO Gastrointestinal Cancer symposium. My name is Katie Kelley, and I’m a GI oncologist focusing in hepatobiliary cancers from here in town at UCSF, and it is my great honor to be standing in as moderator for Dr Neil Love tonight, who is certainly a tough act to follow.

I’d like to start by introducing my colleagues and co-faculty for this event. Speaking first we have Dr Richard Finn, who is a Professor in the Department of Medicine and Medical Oncology at UCLA, the David Geffen School of Medicine. We also have Dr Lipika Goyal, to my left, who is the Director of Gastrointestinal Oncology at Stanford Cancer Center. And Professor Arndt Vogel, who is a Senior Consultant hepatologist in the Division and Department of Gastroenterology at the Hannover Medical School in Germany, so certainly has the most reason to have jetlag of any of us here.

So before we proceed I’d like to acknowledge our commercial support.

So without any further ado we’ll launch into our tonight’s program, which will take an hour and a half; again, cases from the community. Our agenda begins with Module 1, first-line treatment for advanced hepatocellular carcinoma. We’ll begin with a case from a community faculty member and the continue after that with a presentation on first-line therapy from Dr Finn. Module 2 will delve into selection and sequencing of therapies for relapsed and refractory HCC, for which I’ll provide a presentation after, again, we receive a case or a series of cases from our colleagues in the community.

Module 3 will focus on the current and future role of immunotherapy in the treatment of advanced biliary tract cancers. And again we’ll start with a few cases and then hear from Dr Vogel for a presentation. And finally we’ll conclude with a session on the integration of targeted therapy into the management of advanced BTCs, or biliary tract cancers, with, again, starting with some cases and then a presentation from Dr Goyal.

And I’d like to also welcome and thank our colleagues and co-faculty from community centers who treat so many patients and are sharing their expertise and questions with us, and their clinical expertise.

Case: A man in his early 70s with newly diagnosed metastatic HCC receives atezolizumab/bevacizumab — Warren S Brenner, MD

DR KELLEY: So with that we’d like to launch into Module 1.

DR LOVE: Good evening, everyone. I’m Neil Love from Research To Practice and welcome to our real-world case library. Tonight, our panel of oncologists will present cases of hepatocellular and biliary tract cancers from their practices and present relevant management cases to our faculty. We begin with Dr Warren Brenner who is treating a 70-year-old man with advanced HCC responding to atezolizumab/bevacizumab.

DR BRENNER: A 70-year-old patient who was found to have a large right-sided hepatic mass, as well as a large, central big mass that was biopsied and consistent with a hepatocellular carcinoma. This occurred in the setting of a prior history of hepatitis C that had been treated, as well as genetic hemochromatosis.

So patient started therapy with bevacizumab and atezolizumab.

This is a graph of his alpha fetoprotein that declined significantly from over 4,000 to within a normal range after completing 4 cycles of therapy.

So my questions for the investigators are, is there a role for genetic profiling in patients with HCC? Is a biopsy really necessary in patients with clinical HCC and high alfa fetoprotein to be greater than 400?

How are the investigators choosing between atezolizumab and bevacizumab versus dual checkpoint inhibitor therapy, based on the HIMALAYA trial?

Are there any clinical parameters to decide between the 2 regimens such as patients with hepatitis-associated HCC more likely to respond to 1 regimen versus another?

Do all patients with HCC receiving bevacizumab require upper endoscopy prior to initiating therapy?

DR KELLEY: So that’s a great series of questions from this first case. And so I think I’ll triage the question first about the diagnostic question at hand, which is the role for biopsy, and I’ll start with Dr Finn. Is a biopsy necessary in patients with a clinical diagnosis of HCC and high alpha-fetoprotein, greater than 400?

DR FINN: I would say by and large it is not, but we do need to be clear on what we mean by a clinical diagnosis, right? That is a very strictly-defined criteria that must be met. A noninvasive clinical diagnosis requires someone to have cirrhosis. It requires them to have a hypervascular lesion on dynamic imaging, either a triple-phase CT scan or MRI, where there’s rapid enhancement and washout on the delayed phases, meeting LI-RADS criteria. LI-RADS 5, we can say with over 90% confidence that this is liver cancer. AFP really does not come into the diagnostic criteria clinically.

DR KELLEY: Thanks so much, Rich. I think this case also exemplifies the active question we all face in the clinic now that we have 2 first-line combination immunotherapy regimens available, with atezo/bev and now the STRIDE regimen from HIMALAYA. And so DR Brenner’s second question for us, which I’ll direct to Dr Vogel, is how are we choosing between atezo/bev versus dual checkpoint inhibitor therapy in a first-line patient like this, assuming no clear-cut contraindications to either?

DR FINN: So I think that’s a good question, and I think the first point is it is good that we have options today. So in the old days it was only sorafenib, which was good, but not good enough. Now we can decide, and we do not really have head-to-head comparisons between atezolizumab and dual checkpoint inhibition. So I’m going to need to do some cross-trial comparison, which always has some caveats. And I think we cannot focus on 1 endpoint today, but we need to look at the totality of data, and with that we need to make decisions based on a case-by-case basis and then look for comorbidities, for example, liver functions, aim of treatment, tumor burden, and make our decisions. At this point in time I would consider atezo/bev as the preferred regimen for most of my colleagues.

DR KELLEY: Thanks so much, Arndt.

Case: A woman in her late 50s with a history of hepatitis C and chronic kidney disease on dialysis with localized HCC after surgical resection; patient developed metastatic disease 10 months later — Liudmila N Schafer, MD

DR KELLEY: So I think with that we’ll move to our second case to make sure we have time to review it as well.

DR LOVE: Dr Liudmila Schafer is currently deciding on first line treatment of a 59-year-old woman with metastatic HCC.

DR SCHAFER: A 59-year-old woman. She is known with history of hepatitis C, status-post treatment back in 2005. She developed a chronic kidney disease and now she’s on dialysis. So she underwent resection for hepatocellular carcinoma.

Her tumor was in segment 7 about 5 x 6 cm.

And alpha-fetoprotein at that time was about 17,000.

Most of the patients as we treat right now, they just get surgery and then we follow up in surveillance.

She subsequently progressed within 10 months and developed metastatic disease with new lesions in the liver and lymphadenopathy.

Otherwise, in a good performance status, getting dialysis.

DR KELLEY: So I see we have some questions for the faculty. What treatment would we recommend? Again, this is a first-line patient after surgery and then recurrence. So maybe Dr Goyal. And with the important caveat of a comorbidity here. This is a dialysis patient.

DR GOYAL: Yeah. I think that’s an important consideration. Of course in the trials that we have the patients with dialysis were not included. So to be very honest I’d have to look at the different drugs that we have in the first line to see in patient with dialysis what would be safe, whether we could do immunotherapy or do TKIs.

I would also say that if she has liver-confined disease only, even though she has metastatic disease, it would not be unreasonable to consider liver-directed therapy such as embolization.

DR KELLEY: Great, thanks. And I guess Rich, I’ll ask, do you have any other thoughts given the comorbidity profile? One question I guess that came up in the last case that we could also ask here is the role of an endoscopy. Would you make — in a person like this would you want to get an EGD before deciding on therapy?

DR FINN: So as far as the renal function, I don’t know that that would necessarily play into a role. Proteinuria is no longer relevant, I would say, when someone’s on dialysis.

Regarding the endoscopy, this patient, we could probably say with pretty high certainty, they do not have portal hypertension given that they had a liver resection a few months ago, and that would not have happened. With that being said, standard of care would be to get an endoscopy if we’re considering using bevacizumab. Whether that needs to be done before the first dose, I think we can debate that, certainly, in someone who is relatively low risk, which I would argue this patient is.

DR KELLEY: I think that’s where I guess I would say that I tend to be more conservative and err on the side of getting the endoscopy, particularly in patients with progressive disease in the liver because I think we can’t always radiographically or based on platelets or other signals assess how much portal hypertension they’re getting with recurrence. So I probably would err on the side of getting an endoscopy, but I agree with you in occasional cases. I don’t know. I guess maybe a tiebreaker. Arndt, are you routinely getting an endoscopy for everyone or are you making case-by-case decisions?

PROF VOGEL: So I’m a hepatologist, so what do you expect —

DR KELLEY: True.

PROF VOGEL: I can do it myself the next day, so it’s no problem. And I would do it actually, yes. But I agree with Rich. I mean we have also indirect markers of portal hypertension. If we do not see anything on the CT scan, platelets are good, spleen is small, and if endoscopy is not available, I’m not sure whether you really need to delay the treatment here. But in general I would recommend to do the endoscopy. We have not only seen bleeding from varices, as you know, but also from ulcers for example, so it’s always good to do some endoscopy.

DR KELLEY: In a patient with dialysis you always want to think about their blood pressure, of course, as well —

PROF VOGEL: Yes.

DR GOYAL: — as bev can sometimes have increased hypertension. And so how that would impact their ability to get dialysis and their overall safety is something to consider of course.

DR KELLEY: I think also a piece of clinical data that I would be curious to know, in this patient, but also patients in making that decision of atezo/bev versus durva/treme, particularly when there are vascular comorbidities at hand. How quickly is the disease growing? If it’s rather indolent I think sometimes, particularly if I’m worried that there’s more of a risk of bleeding or labile hypertension I might be more inclined to use the STRIDE regimen as a first line, again, in a patient with comorbidities and particularly if we think they have a little bit of a runway for a dual immunotherapy approach.

So I think with that we can proceed with our first-line talk from Dr Finn.

Faculty presentation: Dr Finn

DR FINN: Okay. Thank you very much, Katie, and thank you to Neil Love and Research To Practice for having me. And I have the honor of discussing front-line management, which is that those 2 cases were referring to. And when we think about the role of systemic treatment and where it fits into the Barcelona staging system clearly for a long time we’ve known the clear advanced patients, those that have metastatic disease or vascular invasion.

But increasingly we’re recognizing there’s a group of patients who have intermediate disease, tumor confined to the liver, no vascular invasion, but they have tumor characteristics that suggest they would not do well with locoregional treatment. And in this updated guideline that came out last year they used this word diffuse, infiltrative, extensive bilobar involvement would be patients who are probably better served getting systemic treatment.

And that other group of patients are the patients who stage migrate. So they do get locoregional treatment, TACE or Y-90. That manages their disease for some period of time, but eventually they become refractory. Despite these procedures on imaging they keep developing recurrent intrahepatic recurrences or even develop extrahepatic spread, and these 3 groups are the ones that we’re talking about front-line systemic management.

The pivotal study of atezo/bev I was very blessed to be part of. The IMbrave150 study took patients who were advanced liver cancer, Child-Pugh A liver function like every study we’re talking about, and randomized them to atezo/bev versus sorafenib. It was an open-label study, randomized 2:1, with 2 coprimary endpoints.

Importantly because of bev’s concerns with bleeding patients were required to have an endoscopy within 6 months of coming on, and if they had high-risk varices, high-risk for bleeding, they were excluded. Varices were not exclude. If they were there they had to be managed per the institutional guidelines. And we do know that 25% of patients did have varices at baseline, and you can see that here, and a number of them had them managed at baseline.

This study, a majority of patients came from outside of Asia, and we see here all etiologies of liver cancer are represented, with about 50% of patients having hepatitis B. And this was a group of patients that had main portal vein invasion, VP4, a very poor prognostic sign, which was excluded from several of the studies we’ll talk about.

This was a hallmark study in that it met both its primary endpoints at the interim analyses, and last year it was published. The updated where the true survival with atezo/bev in this population is 19 months, 19.2 months, including patients that had main portal vein invasion, and similar benefit when we look at PFS by magnitude. What’s really most striking about this regimen is the 30% objective response rate, which includes 8% CRs. And like with all IO regimens if you respond you do very well; long duration of response.

And this regimen appears to be well tolerated. If we look at the most common adverse events they’re clearly driven by bevacizumab, and that is hypertension and proteinuria. Certainly immune-related adverse events occur but now with such high frequency. Most of the things that patients notice: diarrhea, hand-foot skin syndrome, anorexia. These occur more commonly and typically with sorafenib. And we know that quality of life with IO regimens has consistently favored those IO regimens.

Now the HIMALAYA study was presented at this meeting last year and subsequently published in NEJM Evidence over the summer and recently got FDA approval, in the past few months. Again, in contrast IMbrave150 they excluded patients who had main portal vein invasion, a group of patients that do poorly. The primary endpoint of this study was comparing a single dose of the CTLA4 antibody tremelimumab with durva versus sorafenib. There was also a single-agent durva arm to compare to sorafenib. Again Child-Pugh A advanced liver cancer.

This study similarly took patients, a majority came from outside of Asia, a little less hepatitis B, about 30%, and like in all studies there’s groups of patients that are intermediate, the Bs. This study met its primary endpoint, improving overall survival with the STRIDE regimen. Median survival 16.4 months in this population of patients who don’t have main portal vein. Hazard ratio’s 0.78, a clear benefit of superiority to sorafenib.

Durvalumab met its noninferiority endpoint. Here you see a hazard ratio of 0.86. Upper limit of the confidence interval is 1.03. Single-agent durvalumab did not get FDA approval, unlike the combination.

PFS, really no difference between the regimens, and response, as we saw with other IO combinations, 20% objective response, single-agent durvalumab 17%, and single-agent sorafenib 5%. And if you respond you respond very well. These are long responses.

Here looking at adverse events, all-cause adverse events very similar between the arms. Treatment-related adverse events actually higher with sorafenib, including Grade 3/4 events. So what you can see is that the IO regimens tend to be well tolerated. Here you see any serious treatment-related adverse event was higher with T300 + D as compared to single-agent durva, as compared to sorafenib.

Now immune-related adverse events are the unique toxicities of these IO agents, and what’s quite striking, even though 1 dose of tremelimumab improves survival it also essentially doubles the immune-related adverse events. You can see here all-grade events with the combo 36% versus 17% with durva alone. And the use of steroids essentially was doubled, from 10% to 20%.

Tislelizumab, this is a unique PD-1 inhibitor. It’s an IGG4 PD-1 inhibitor which was designed to not engage Fcγ receptors, which that engagement can downplay the anti-PD-1 response. At ESMO we saw the results of RATIONALE-301, which was tislelizumab versus sorafenib, again in patients without main portal vein invasion. This was a noninferiority study, and like other single-agent PD-1 inhibitors we see a hazard ratio of 0.85. That’s very consistent, very consistent message with single-agent PD-1 inhibitors. Hazard ratio 0.85. The upper limit of the confidence interval was 1.019, which is less than the noninferiority margin.

Single agent, 14% objective response rate. These were very long duration of response. Actually, tislelizumab had a median duration of response of 3 years if you responded. That is quite long for a single agent. But no differences in PFS, maybe even a little longer with sorafenib.

The safety, very clean safety profile related to immune-related adverse events. The use of steroids was about 10% or less.

Now obviously if you can’t get an IO regimen for the various reasons there are 2 TKIs that have been shown to improve survival, I should say sorafenib versus placebo, that was superior. Lenvatinib versus sorafenib was noninferior. And you can see here in the pivotal REFLECT study median survival was 13.6 months.

Recently at ESMO we saw the results of LEAP-002, len/pembro versus len/placebo. This was one of the first double-blind placebo-controlled studies in the modern age, because most of them are open label IV versus PO. Here we see survival with lenvatinib as a single agent, 19 months, but keep in mind, this 19 months is not the same population as IMbrave150 because this study excluded main portal vein invasion.

So in conclusion we’ve made a lot of progress in improving survival for our patients. Introduction of IO is now front line. In response to one of the questions, we have no biomarker to select patients, no clinical marker really to select patients. If you’re not a candidate for a doublet consider TKIs or even single-agent IO, and studies evaluating these in earlier-stage disease are ongoing. And in fact, 2 days ago we heard a press release of the IMbrave050 study, which we’ll wait to see those results, but suggests that atezo/bev improves recurrence-free survival, delays recurrence after resection. And so again, that case we heard about at the beginning perhaps in the future would get adjuvant atezo/bev and thank you very much.

DR KELLEY: Thanks so much, Rich. And just direct folk’s attention, this is the press release that came out on the 18th for a randomized Phase III trial atezo — called IMbrave050 that randomized patients after resection or ablation for high-risk HCC to receive either atezolizumab plus bevacizumab for 1 year or close monitoring. And they reported positive recurrence-free survival endpoint results, which was the primary endpoint, 2 days ago. So stay tuned, obviously, for a very big presentation. It’s the first adjuvant trial to report positive results in contemporary HCC history.

Case: A man in his mid 60s with metastatic HCC and a medical history of hepatitis C develops interstitial lung disease on atezolizumab/bevacizumab — Priya Rudolph, MD, PhD

DR KELLEY: So with that we will move on to Module 2, and this is a topic of selection and sequencing of therapies for relapsed and refractory HCC, and I’ll be presenting the presentation, but first we have a case to start with.

DR LOVE: Dr Priya Rudolph is managing a challenging case of a 65-year-old man with metastatic HCC.

DR RUDOLPH: This is an unfortunate case. It actually happens to be the father of one of our vendors.

A 65-year-old male who has a history of hepatitis C.

I started him on treatment initially with atezolizumab and then added on bevacizumab to the regimen.

Four months later, he started developing dyspnea on exertion. I did a CT scan, which showed ground-glass opacities. He was COVID negative, and upon further questioning he tells me that he had been cleaning his boat and possibly inhaled some strong chemicals.

However, he was on atezolizumab, which certainly can cause interstitial lung disease, so I decided to hold it and put him on a steroid taper.

Is there any data to suggest that continuing single-agent bevacizumab has any efficacy, and can we just continue that for a while?

Really the question for the future is, if he has progression what are our next options.

DR KELLEY: So that’s a great question. This is — I think a lot of face this challenge when dealing with immune-related, or potential immune-related adverse events in the clinic, of discerning is this a dreaded complication of immunotherapy or is this a complication of some other event, like in this case the inhalant exposure or rash for other reasons or diarrhea for other reasons. And so I think we can’t always know. But in a case like this, I’ll speak for myself and then poll my co-faculty here, I think I would hold the atezo for the moment and engage the patient’s response to steroids, see how quickly he responds to get a sense of how responsive he is to steroids, but also get a consult from a pulmonologist to guide — to help me in the navigation of the decision making.

And to the question of continuing bevacizumab, I would continue that drug because in the IMbrave150 trial the alternate agent was allowed to be continued in cases of the second drug being held for toxicity reasons. So I think I would continue bevacizumab given the patient was doing well, in parallel with a workup with a pulmonologist. And again, depending on the patient’s response to steroids and how confident the pulmonologist is that it’s an inhalant etiology versus immune therapy, if it’s not too severe there could be a very cautious rechallenge, but I think I only would do that in consult and in collaboration with a pulmonologist, a colleague who’s an expert in that disease — or in that organ system to make the decision with me.

So I don’t know if any other colleagues want to jump in. Lipika, you’re nodding.

DR GOYAL: Yeah, no. I totally agree with you, Katie. I would say that it’s really helpful to have our colleagues in pulmonology help with pneumonitis and gastroenterology help with colitis, et cetera. I always think when people have like thyroiditis, colitis, dermatitis, I’m much more comfortable rechallenging. With pneumonitis I also agree with Katie that I would really consider rechallenging this patient given how active atezo/bev is if they only had, for example, Grade 2 pneumonitis that the pulmonologist felt was related to atezolizumab. When people have Grade 3 pneumonitis I’m much more cautious and often do not rechallenge. With Grade 3 colitis there are some data that it’s not unreasonable to think about rechallenge with an immunotherapy, sometimes with an immunomodulatory agent onboard, but certainly would monitor very closely with the help of our colleagues.

DR KELLEY: Rich, what do you think?

DR FINN: No. I agree with that. This isn’t unique to liver cancer this question. I mean we’ve all been using IO in every disease, and these things happen. So I think pulmonary — if you’re thinking about steroids then it’s probably time to get pulmonary involved.

DR KELLEY: Okay. And Arndt, any irAEs that you absolutely would never rechallenge, or any comments on ones that you feel very uncomfortable rechallenging?

PROF VOGEL: I mean as Lipika just said, it really depends on the severity and how the patient was affected by it and how he or she responds to steroids. So yeah. I think it’s really rare in HCC, right, specifically with atezo/bev. I mean I can hardly remember a patient which I did really completely stop forever. It’s different when you use dual checkpoint inhibitors, but with atezo/bev I think it’s quite feasible.

DR KELLEY: I think there is some particular — like myocarditis, for example, where I wouldn’t rechallenge, but I think, again, this is particularly with an alternate etiology on the table and low-grade it makes sense.

PROF VOGEL: We had 1 case with autoimmune hepatitis, actually. And he developed a flare of the hepatitis and there we did decide not to rechallenge.

Cases: A man in his late 60s with post-transplant recurrence develops metastatic HCC and receives lenvatinib followed by cabozantinib followed by regorafenib and a man in his late 70s with metastatic HCC after treatment with atezolizumab/bevacizumab, lenvatinib and currently cabozantinib — Lionel A Kankeu Fonkoua, MD and Susmitha Apuri, MD

DR KELLEY: Great. Well I think we have another case to review now.

DR LOVE: Dr Lionel Fonkoua is managing a 68-year-old man with metastatic HCC.

DR FONKOUA: So this case is actually a patient that had actually a transplant years ago for HCC, multifocal HCC in the background of alcoholic cirrhosis, and was doing fine on immunosuppressive therapy, but had a post-transplant recurrence.

Started on lenvatinib.

And then eventually progressed, switched to cabozantinib and then progressed on that after about 4 months of therapy. He was placed on regorafenib. In a patient like this would you consider atezo/bev?

And then the second question is there’s been some data suggesting that activating mutations in β-catenin actually confer resistance to immune checkpoint inhibition. Would that factor in this setting, in a patient who you already have reservations?

We actually had a discussion about that, and he wanted to remain aggressive, but I just didn’t feel comfortable knowing.

Graft failure rate of about 30%, 40% in this patient population.

DR LOVE: A more typical case of metastatic HCC is a patient of Dr Susmitha Apuri, a 79-year-old man currently feeling well on third line cabozantinib.

DR APURI: I thought he was progressing.

He’s on cabozantinib right now.

Would he be a candidate for pembrolizumab?

Is there any role for alpelisib and fulvestrant?

We are looking at somebody who currently has an ECOG of 0. He’s back to his woodworking.

Every time he walks in he says I feel great, don’t take me off of the cabozantinib.

I would love to know if he would be a candidate for ipi/nivo.

DR KELLEY: Great. So these are 2 great cases. One is a very esoteric case in a post-transplant patient, the second a more bread-and-butter case of first-line, second-line, third-line therapy and then what. So I think starting with the first — the first patient, who’s a post-transplant patient, I think Arndt I’m going to go with — go to you for this one as our resident liver expert. What is your stance on immune checkpoint inhibition post-transplant population?

PROF VOGEL: Yeah. I think it’s a very relevant question. Certainly checkpoint inhibitors would not be a preferred first-line treatment. I would consider it a contraindication. But to be honest, we have already treated patients after transplantation with checkpoint inhibitors. And there’s also some literature out there that it is possible.

But you always need to be aware that around 30% will develop severe rejection, which can lead to death. And therefore I think — I mean of you have other options I would not go for a checkpoint inhibitor. If there’s no other option, and you included a chemotherapy, local therapies, and whatsoever, we have actually used checkpoint inhibitors, and I think it is possible, and our — I mean we have 2 patients who did not develop rejection.

In respect to the other question about β-catenin mutations, I think it’s interesting. So we thought it would mediate resistance to checkpoint inhibitors, but I think it’s not so clear. There are actually data from the IMbrave study and also from the CheckMate study that these patients might be more sensitive to sorafenib. So I think maybe we can also consider sorafenib at some point in this patient.

DR KELLEY: Yeah. I’ll chime in as the second-line speaker ahead, I think in this case I probably would start with sorafenib first, that is — as Arndt mentioned, there is some data from the CheckMate 459 study of nivolumab versus sorafenib which showed that β-catenin mutations were favorable in the sorafenib arm and really no impact or no significant effect one way or the other in the nivolumab arm in that randomized context.

Regarding immune checkpoint inhibition post-transplant, I think the literature we have so far suggests that the rate of fatal rejection is as high if not a little higher even than the rate of likelihood of shrinkage. And so for a palliative therapy I have a really hard time with that, though I certainly agree with Arndt’s point that if a patient is end-stage and is otherwise super fit it’s a hard conversation to consider.

I think my sense on that would be generally almost always no, but the only times I would consider it is if there’s a very actively involved expert transplant center that has experience in managing the complications of rejection post immune checkpoint inhibitor therapy. So it’s not something that I would ever consider unilaterally, and I would say it should be something only done after a referral to the transplant center, which needs to be a very experienced transplant center with oncology patients. And again, as a rule it would be a no for me.

Rich, do you have any other —

DR FINN: I have done it in 2 patients now, but to the point that you described. The patient who’s gone through lenvatinib, sorafenib, regorafenib, ramucirumab, cabozantinib, and is still fit, and we have a discussion, right? The interesting thing is we never take them off their immunosuppression, right? The hepatologists are like don’t take them off, but we’re kind of putting on the accelerator and the brakes at the same time. So I’ve never seen benefit, but I didn’t see rejection. But again, never would that be a first-line option for me in a post-transplant.

DR KELLEY: Any other thoughts Lipika?

DR GOYAL: Yeah, no. Agree with all. Like a last resort, but certainly something that’s worth considering if someone has been through all the regular therapies. And I also would encourage that every once in a while we have clinical trials that allow patients with transplants. Some of the FGFR4 inhibitor trials have allowed patients with transplant, so it’s certainly also worth considering clinical trials if it’s on the eligibility criteria.

DR KELLEY: Yeah, I think that’s a great point, that this would be a patient I would definitely preferentially refer for a clinical trial before immune checkpoint inhibitor given the risk and the liability and the activation energy for both I think favors trial.

So the second patient. He had atezo/bev, then lenvatinib, then currently third-line cabozantinib, and then a somatic PIK3CA mutation. I think for a case like this, this is also someone where we don’t have — we’re in uncharted waters with respect to comparative randomized trial data. We don’t have data in the fourth-line setting. I think if there’s been enough time elapsed, and particularly noting that perhaps the immune microenvironment may have some plasticity over time, this is somebody where if they’re very fit I would potentially consider rechallenging with an immune checkpoint inhibitor.

Sorafenib is also another TKI to consider. So nivo/ipi in a 79-year-old who’s fit and ECOG 0 I think is a reasonable choice with close monitoring and careful monitoring of the patient in follow up. I think sorafenib is another choice that I would consider. I don’t think I would use a PIK3CA inhibitor in this context.

So actually I’ll pick on I think Lipika again. Any other thoughts?

DR GOYAL: Yeah, no. Agree with that. My first choice would be considering ipi/nivo if it’s a fit 79-year-old. And then PIK3CA mutations sometimes are clonal, sometimes are subclonal. I’d be curious to know if that was found on tissue or in blood. And then there are activating PIK3CA mutations and mutations that are not activating. Again, it’s worth considering a clinical trial if it is an activating mutation that is clonal. But again, a second resort.

DR KELLEY: Rich and Arndt, would either of you use a PIK3CA mutation, assuming it’s activating — or inhibitor?

DR FINN: Yeah. I think you’re reaching there, right? It’s context dependent. This is a validated target in ER-positive breast cancer, but I don’t think in any — despite it being widely mutated drugging it has never correlated with any activity, I think, outside of that indication.

But I think the ipi/nivo question is important because what it reflects is we want to harness IO responses, right? And even though someone progresses on atezo/bev or another IO regimen, does the partner matter, right? Does a TKI/IO reverse something that was missed by bev? Does CTLA4/IO capture something that was missed by one of the other regimens? And there’s hints of that in small series, but for someone who’s fit, who’s gone through other things, I’ve gone back, and I’ve seen responses with ipi/nivo actually, after prior IO.

DR KELLEY: I have, as well, and I think the melanoma data reassures us that it’s possible, and I think also the potential for neutralizing antibodies on atezo, though rare, and we don’t fully know their impact in real world, I think also argues for considering it. I don’t know. Arndt, did you have anything you wanted to add?

PROF VOGEL: Just I mean we have clinical trials, like IMbrave 251, that we can look at this in terms of a prospective trial, and I think this concept of IO beyond progression and all the questions, I mean they can be addressed in a clinical trial, and we have these clinical trials so we should really include our patients.

And regarding PIK3CA, I think in TAPESTRY, for example, they require 2 mutations. That’s why I think with a single mutation in a solid tumor, except breast or anything else, I probably — we wouldn’t do it, no.

DR KELLEY: Great. Thanks so much.

Faculty presentation: Dr Kelley

DR KELLEY: Well that brings us to the didactic section that I’ll be giving for this Module 2, which is selection and sequencing of therapies for advance HCC second line and beyond.

So first I’ll go over the prevalence of use of subsequent therapies in advanced HCC and then cover our currently available second-line therapies, ranging from agents with FDA approval and after sorafenib, to those — how do we incorporate those after immune checkpoint inhibitor-based combinations, the factors we use in selection and sequencing of second-line or later options, and other promising strategies on the horizon.

So a meaningful number of patients now receive greater than or equal to second-line therapies. This may be due to our more effective first-line agents, better supportive care, like hepatitis B and C therapy, and possibly stage migration. We’re treating patients earlier. And so you can see the IMbrave 150 data that almost 50% of patients, when you combine both arms, went on to receive at least 1 and at least 30 to 50 went on to receive a second line of therapy thereafter. And then even a meaningful number get third line when you consider the number of patients in the HCC population.

And we see similar data from the randomized — large, randomized Phase II data set from Study 22, which was a predominantly second-line trial, and we see again a significant number, greater than 40% across arms of that trial combined, when on to receive a third line. So again, this is not a trivial amount of patients.

And this is the BCLC algorithm that Rich showed in the first-line setting, and I’m just going to highlight here the systemic therapy options in second line, where it’s very clear that we really don’t have Level 1 evidence for what to do after sorafenib. After sorafenib we have Level 1 evidence for regorafenib, cabozantinib, ramucirumab if AFP is elevated, but after atezo/bev, durva/treme, or lenvatinib or durvalumab we are really operating — flying blind right now.

So just to review where we are in beyond first-line context, we have antiangiogenic therapies with USFDA approval and Level 1 evidence after sorafenib, including, as I mentioned, rego, cabo, and ramucirumab for high AFP greater than 400. And then we have accelerated approval based on Phase I and II data from pembrolizumab, as well as nivo plus ipi. Again, all of these are after sorafenib. And as Rich mentioned in the first-line setting, it would be remiss of me not to mention all of these are in Child-Pugh A populations even though as we get later in our therapy not everyone is Child-Pugh A. So it’s another need area.

So focusing on the first group I think we’ve all — here we go — we’ve all seen these topline primary endpoint results, but just for a refresher we have the RESOURCE trial of regorafenib versus placebo showed improvement for regorafenib in patients who had tolerated prior sorafenib therapy with a dose of at least 400 mg per day for at least 20 of the past 28 days. Cabozantinib, the second Kaplan-Meier curve you see, also improved survival over placebo in both the second and third-line context population after sorafenib and up to 1 other line of therapy, so had a little more permissive entry criteria. And then ramucirumab improved OS in the REACH-2 trial when excluding to patients with high AFP for eligibility — or selecting for those patients.

Now the question that comes up, what do we do with these multikinase inhibitors after atezo/bev. There is very limited prospective data for treatment selection after first-line immune checkpoint inhibitor therapy. I’ll draw you attention to a retrospective analysis, so again with all the hazards thereof, of 71 patients treated with a second-line TKI, sorafenib or lenvatinib, after first-line atezo/bev. And it showed for either sorafenib or lenvatinib a median PFS of around 3 months, median overall survival 14.7 months, which is, again, similar to what we’d expect in any second-line context, so reassuring that their efficacy is relatively preserved after atezo/bev just as it is after sorafenib. And lenvatinib had the better performance with PFS of 6.1 months. So again, this is retrospective but reassures us that the numbers that we’re seeing are quite similar to what we expect after sorafenib.

Similarly, there’s a systemic literature review of cabozantinib after prior ICI therapies across tumor types which shows very similar outcomes of efficacy, as we expect after more conventional first-line therapies, which was studied in the first — second-line setting originally.

So beyond first line we have also some agents with accelerated approval based on Phase I/II studies. As I mentioned, pembro and nivo/ipi, and both of those have been studied also after sorafenib. So the pembrolizumab is an accelerated approval originally based on KEYNOTE-224, which was a single-arm study that showed a really encouraging response rate, but in the follow up Phase III trial, KEYNOTE-240, the overall survival did not show a significant improvement, though secondary endpoint showed trends towards improvement with pembro, so it was a near miss. An ensuing study, KEYNOTE-394, in an Asian population did meet its endpoints and showed significant survival benefit, which I’ll show in the next slide.

And then the bottom waterfall plots from nivo/ipi showed deep and durable responses to nivo/ipi in the CheckMate 040 trial, which looked at 3 different dosing regimens. 30% response rate and a prolonged duration of response and actually a quite encouraging overall survival of over 25 months — or 28 months in Arm A, nivo 1/ipi 3 given 4 doses every 3 weeks followed by nivo maintenance. And so based on the strength of these robust responses this regimen also garnered an accelerated approval.

So KEYNOTE-394 was the follow up study after KEYNOTE-240 was negative. KEYNOTE-394 was conducted, as I mentioned, in a predominantly Chinese population. And you can see the hepatitis B status was in almost 80% of patients in both arms. And here we see that the survival met its primary endpoint with hazard ratio of 0.79, significant P-value, and median overall survival of 14.6 versus 13.0 months, and responses were significantly favoring the pembro arm, 12.7% had responses in this second-line population versus 1.3 for placebo.

And a meta-analysis looking at the 240 plus 394 studies together reinforced that these are consistent with each other even though, again, 240 did not quite meet significance.

So there’s really no clear role for how to use immune checkpoint inhibitors themselves after another prior checkpoint inhibitor, as we discussed with the last case. There is some data in other tumor types that adding in a CTLA4 can restore response, as we talked about with the case we just discussed.

And I think that this brings us to some factors to consider, which first-line therapy was received? Are there contraindications to immune checkpoint inhibitors in the second line or contraindications to antiangiogenic therapies? And are they eligible for ramucirumab? And what’s their liver function?

So right now I think our decision making about which drug to use in the second line really depends first and foremost on what did they receive in first line. If they received sorafenib or a tyrosine kinase it’s very easy. We can decide to use our new first-line regimens or a Level 1 evidence drug after sorafenib. If they have had prior atezo/bev then we are really flying without Level 1 evidence and making decisions based on comorbidity profile and their response to prior therapy.

So the current NCCN Guidelines highlight here for second-line subsequent therapy. You can see at the bottom that there are — it’s pretty agnostic to what patients received in first-line, showing all of the antiangiogenic regimens, with rego, cabo, ram, lenvatinib, and sorafenib being listed as reasonable options after a first-line regimen. And then nivo/ipi is listed as a consideration with Level 2B strength of evidence, as is pembro. But again, we don’t know how to use those after immune checkpoint inhibitors first line.

And so here’s a summary. After our current second-line regimens, including antiangiogenic agents, rego, cabo, ram, and then I think in patients who received atezo/bev or durva/treme first, our conventional first-line options can be bumped on as a second-line choice. So I think a lot of us use lenvatinib as the second-line regimen in part based on the retrospective data I showed after atezo/bev for the tyrosine kinase inhibitors. And all current FDA-approved treatments have been studied after first-line sorafenib.

And then I want to show you — I think Rich mentioned and Arndt mentioned the Phase III IMbrave251 trial, which will help us — I think will be the first trial that gives us a little bit of insight into what to do in this context. This is the schema for that trial. The IMbrave251 trial randomizes patients after progression on atezo/bev to continue atezo and then add a historically first-line TKI, so either lenvatinib or sorafenib, versus switching to lenvatinib or sorafenib alone. And the physician’s choice of which TKI to use.

And this trial, I think, will give us a lot of information about response rate to the TKIs and time to event endpoints for the TKIs alone and whether there’s any benefit to continuing immune checkpoint inhibitor after progression.

There’s a lot of new horizons — new agents on the horizon, including new combinations and some approaches to combine other anti — immune checkpoint inhibitors or other immunotherapy modulating agents to see if we can harness the immune response in a better way.

And with that I will close.

Case: A woman in her mid 70s with metastatic cholangiocarcinoma who received first-line gemcitabine/cisplatin with durvalumab — Dr Brenner

DR KELLEY: And I think that brings us to Module 3 next, which is current and future role of immunotherapy for biliary tract cancers. So we’ll be changing to a different primary liver type, biliary tract cancers, starting with several cases.

DR LOVE: Dr Brenner is managing a 75-year-old woman with metastatic cholangiocarcinoma who received the first line therapy studied in the Phase III TOPAZ trial, gemcitabine, cisplatin and durvalumab.

DR BRENNER: This is a 75-year-old patient who had metastatic cholangiocarcinoma from an intrahepatic cholangio.

Who received standard therapy with gemcitabine and cisplatin together with durvalumab. Completed 8 cycles, but, unfortunately, progressed shortly after completing her chemotherapy.

This is a picture of her intrahepatic tumor. And sitting just above that some of the disease on the surface of her liver.

My questions for the investigators are, are there subtypes, such as gallbladder cancer versus intrahepatic versus extrahepatic cholangiocarcinoma who may respond differently to these agents, particularly checkpoint inhibitors?

Is there any differential effect based on PD-L1 and tumor mutation burden? And what are the management options for patients with progressive disease who have no targetable mutations?

You know what you should ask the investigators, because I think this is happening a lot in the community. There’s a lot of liquid biopsies and a lot of the tumor mutation burden readings definitely come back higher on the liquid biopsies.

I get a lot that are 15, 18, 22. And the original pembro approval was based on, I think it was the Foundation TMB.

So it would be interesting to ask them, do they use tumor mutation burden with liquid biopsies to use potentially pembrolizumab with high tumor mutation burden?

DR KELLEY: Great. So that’s a very good representative case. So just to summarize, it’s a 75-year-old lady with an intrahepatic cholangiocarcinoma and primary metastatic disease who started treatment under the care of Dr Brenner with gem/cis plus durvalumab following the TOPAZ regimen, and completed her up-front 8 cycles, or 6 months, but then progressed. And we saw pictures. She has a pretty bulky multifocal intrahepatic tumor and some metastatic diaphragmatic implants.

And so I think the first set of questions that we can ask, and I think I’ll direct these to Arndt because I think you may have presented some data on this at World GI Congress, are there subtypes or does the anatomic subtype, such as gallbladder versus intrahepatic versus extrahepatic, show a differential response in TOPAZ-1?

PROF VOGEL: Yeah. So I think it’s an interesting question. And first of all, I mean it’s a post hoc analysis, so we need to be careful with that. And to be honest, the gallbladder group responded poorly, but this was a very small subgroup which was mainly — which mainly recruited in South America, whatever that means. So I think we need to be cautious. At the moment I would not base my decision making on anatomic location, but I completely agree these are very heterogeneous diseases, and we need to look more carefully at this. But at this point in time I would not make a difference between intrahepatic, extrahepatic, or gallbladder.

Having said that, in the TOPAZ study response was a little bit lower in the gallbladder group.

DR KELLEY: Great. Thank you.

I think the next question we should field is the question about biomarkers of response. And Dr Goyal, I’ll as you this, PD-L1 and TMB, particularly liquid biopsy TMB, in decisions about pembro.

DR GOYAL: Yeah. Dr Brenner is absolutely right that on liquid biopsy the TMB is higher than it is on tissue, and there have certainly been studies showing that, that there’s a bit of an overestimation using liquid biopsy. And so I personally do not use the TMB on liquid biopsy to make determinations around pembro for high TMB.

And PD-L1, there have been a series of studies looking at PD-L1 expression, the nivolumab studies, pembrolizumab studies. It’s not an obvious biomarker that definitely predicts response to immunotherapy. So if someone is — has a PD-L1-positive tumor, they don’t have alternative biomarkers, they don’t have MSI-high tumor or high mutational burden, it’s not unreasonable to think about doing a checkpoint inhibitor in that setting if they have high PD-L1. But of course we have FOLFOX and nal-IRI/5-FU for second-line therapy also.

DR KELLEY: Great. Thanks so much.

Case: A man in his early 70s with localized cholangiocarcinoma deemed potentially resectable with tumor response — Jeremy Lorber, MD

DR KELLEY: I think that brings us to our next case.

DR LOVE: Dr Jeremy Lorber also used the TOPAZ regimen, but in a 72-year-old man with localized, but borderline resectable disease.

DR LORBER: 72-year-old man. His history is notable for Hodgkin lymphoma that was treated in his 30’s with chemotherapy and mantle radiation. He was then diagnosed presently with intrahepatic cholangiocarcinoma after presenting with abdominal pain and fevers.

Based on the location, the surgeon deemed him not a surgical candidate for the time being, but potentially in the future. I started him on durvalumab, gemcitabine and cisplatin regimen.

He’s thus far had a partial response to therapy. He’s tolerated it very well. Some minor cytopenias, but nothing that’s required delay or dose modification. And based on discussion with the surgeon again after the new imaging, he’s still not a perfect surgical candidate and the surgeon ideally would like further response before considering resection. So it was decided that he would continue on the systemic therapy, but would get radiation therapy to the lesion for the time being and see if that helps prolong control or get him a response.

DR KELLEY: Great. So our questions for the faculty for this patient. Any suggestions after this patient has had a partial response to gem/cis plus durva per TOPAZ-1? Has locally advanced disease, may be a potential surgical candidate, considering local radiation. So maybe I’ll turn to Rich next. What would you do in this case? Would you do radiation or more chemo?

DR FINN: Yeah. These are tough cases. And location, location, location. If this is a hilar tumor, a Klatskin tumor, transplant could even be considered for these patients. That is being done in centers. And in that context radiation to the hilar region is typically done before transplant to help control the tumor.

For an intrahepatic cholangiocarcinoma, these are more typical solid adenocarcinomas. You get some response. Maybe the tumor is too big or in a location that the surgeon can’t operate. Sometimes the surgeons will request interventional radiology to do an embolization to the tumor in that lobe of the liver to help hypertrophy the other side of the liver to make them more of a surgical candidate, to give them more wiggle room for resection.

I think locoregional treatments. If it wasn’t too big, even an ablation. We can’t cure this patient without surgery. Our goal is to improve survival. But if they’re still in the realm, and they’re young and motivated to get to surgery, I think an effort, whether it’s external beam or something else, to try to make them resectable is worth considering. If they’re marginal surgical candidates maybe systemic treatment is just a better way to go.

DR KELLEY: Arndt, what do you think?

PROF VOGEL: So I completely agree. I mean I think it really depends a lot on the expertise of the surgeon, so searching the surgeon can also help sometimes. But otherwise I agree. I mean if you get a certain response with chemo and IO, I mean at some point the tumor will most likely not shrink any further, and the question is can we add local therapy. And I think radiation is meaningful. Y-90 you could do. Portal vein ligation to induce hypertrophy could be meaningful.

So I think for this liver-limited disease we need to be openminded to go to a multidisciplinary tumor board and see what we can combine to bring patients to surgery, or as Richard said, to transplantation maybe even.

DR KELLEY: I think I would just add to this that this patient was responding at the time of the initial restaging, and if the patient has had a nice brisk response and is only halfway — maybe 3 months into their therapy, then I would definitely vote to continue. I would only really pause and switch if it’s plateauing or there’s a lot of toxicity.

I guess the one other comment just for a broad audience is that when making decisions about a possible surgical candidate who’s anatomically complex, any decisions about radiation or Y-90 need to be taken in a multidisciplinary fashion and making sure that the surgeon and the IR doctor and the surgeon and the radiation doctor are in very careful communication so that a key part of a future anastomosis, for example, doesn’t get radiation that precludes the future surgery that was intended or that a vessel doesn’t get embolized in anticipation of Y-90 that then is needed for a surgery that was the whole point of it. So this has to be a very multidisciplinary discussion because it’s pretty complicated.

So with that I think we’re at Arndt’s talk. Part 3.

Faculty presentation: Prof Vogel

PROF VOGEL: Yeah. Thank you very much. So I would like to discuss with you the role of immunotherapy in BTC, and some of the points we have already addressed before. We have anatomical heterogeneity, which is important to consider, intrahepatic, extrahepatic, hilar, gallbladder, even — and we also have heterogenic — genetic heterogeneity. And we just recently published in Journal of Hepatology a quite extensive analysis of more than 6,000 intrahepatic cholangiocarcinomas, and here we provide really very I think clear data on the frequency of the druggable alterations we can expect in BTC, and for example the FGFR2 fusion is around 10%, IDH1 mutation is around 15%. So this is a very detailed analysis, and overall you can expect in around 40% of patients druggable genetic alternations.

Therefore, I think it’s really important to perform NGS, maybe not so much to look for TMB, but for some of these alterations which we also have included in the just recently published ESMO Guidelines. As you can see here, we really need to decide what is — in which category does our patient fall; early stage, locally-advanced as we have just discussed. We can also think about local therapies, also included here, and then we always need to go back and ask the surgeons whether we can do surgery.

In the more advanced setting NGS profiling is essential. This is highlighted here. Durva I would consider is the standard of care in combination with gem/cis. And then we have added here the drugs that have already been approved by FDA, and some of them also by EMA.

In terms of biomarkers, also something we have briefly discussed, I mean they are very rare for IO. So TMB is in only around 4% of patients, and MSI around 1%, and roughly 40% of patients with TMB high are actually MSI. So this is a rare population. Usually in BTC we have a low TMB of just around 1.7, so if it’s above 10 I think it would be worthwhile to think also about MSI status.

Otherwise, having this low prevalence of these maybe established biomarkers for immunotherapy the majority of BTCs can be considered as cold tumors. There are some multi-omic classifications, and when you look at these different classes, and I don’t want to go into all the details, around 10 to 30% can be considered as hot tumors, which is maybe in line with the data we have seen for IO monotherapy in BTC, like the KEYNOTE-158 study, response rate below 10% for PD-L1 positive cases, indicating that BTC is not really a good tumor for IO monotherapy.

So we need better biomarkers or we need combinations, and they are a lot of combinations that have been tested. Some of them very well published, like here the study with a MEK inhibitor, for example. I mean very interesting preclinical data suggesting we can get synergistic efficacy when we add MEK inhibitor to a checkpoint inhibitor. This trial was actually positive by definition, but when you look at the PFS both for the single or combination arm it’s moderate and probably not really a breakthrough to follow-up on.

Other combinations that have been tested, lenva/pembro. Richard has already mentioned the LEAP study in HCC, exciting survival data, unfortunately a negative study. So here in BTC similar data. One study from Asia I think response rate 25%, promising. Median overall survival 11 months. Or dual checkpoint inhibition here with nivo/ipi. Again, side effects of course need to be observed, and a response rate of 23%.

So combinations are probably the way to go, and one of the most convincing studies that has been published from South Korea, the Medi-Treme study, in which gem/cis in combination with durva or durva/treme, has been analyzed, and the waterfall plot I think speaks for itself. There seems to be activity specifically in this Asian population. And shortly after that 2 trials were initiated, one was the TOPAZ study, which looked at the efficacy of durvalumab in BTC, and the other one was the KEYNOTE study, which we will most likely see — the results we will most likely see this year.

TOPAZ has been published in the New England Journal of Evidence just last year. And I think in the study we really have to acknowledge that this was only the second positive Phase III trial in the first-line setting and the first positive IO study. We do see a significant survival benefit, and I think in all countries where durvalumab is reimbursed it has been included in the treatment algorithm very fast. PFS slightly improved. Response rate slightly improved. So there is clearly a benefit. Of course we still can do better.

Some subgroups have been looked at. Race, for example, ethnicity. There seems to be like a slight trend for a better outcome for Asian patients, but this is not really significant. Post hoc analysis are all of these analyses I’m just showing you. We already just briefly touched on the question about tumor location, and here you can see the gallbladder patients, which do have a hazard ratio of 0.94, but again, this was a very small subgroup of patients.

In terms of tolerability, I think the tolerability of adding durvalumab to gem/cis is really very well tolerated. Of course we do see an increase of immune-related side effects in 20% of the cases for example, thyroid side effects, which can be easily managed. So toxicity is clearly no problem.

The data have been updated. The survival benefit is consistent, and if you’d like we can see here this trend for long-term survival in BTC, 2 years. As you know, the median overall survival is rather short, so adding a checkpoint inhibitor to chemotherapy is, in my view, clearly meaningful.

So in respect to other biomarkers we have to acknowledge that we do not have a good biomarker. PD-L1 was obviously an obvious candidate to be a biomarker, but as you can see here the different approaches to look at PD-L1 expression, TC score, CPS score, or here in BTC the TAP score, have been looked at. There’s always a slight trend to higher efficacy in patients that do express PD-L1, but it’s not good for decision making yes or no, and in clinical practice I think you do not really need to do it for deciding on durvalumab or not.

How can we make progress? This maybe briefly to show you, a study that we have published last year, the IMMUCHEC study. Complicated design, 5 arms, where we have looked really at different combinations of chemotherapy, durvalumab monotherapy or in combination with tremelimumab. And actually still in both doses the 4 x 75 mg and 1 x 300 mg, as in the HIMALAYA study. The study was negative. There was no strong signal for overall survival. But maybe to highlight here Arm D, the arm that we combined gem/cis with durva and single-shot tremelimumab, had the longest survival of 22 months compared to 16 months in the control arm, which was also already very long, but still there was a trend to a better outcome, indicating that there might be some role for CTLA4 not only in hepatocellular carcinoma but also in BTC, which needs to be explored, of course, in further studies.

What else can we combine? I think the combination of targeted therapy with checkpoint inhibitors could be interesting. Data we have seen from this classification I have mentioned at the beginning suggests that those tumors that harbor druggable alterations like IDH2 — or IDH1 or FGFR2 fusion are more of the non-inflamed cold tumors, and the question is to which to extend these driver alterations to really affect the tumor microenvironment. And when we inhibit the driver maybe we can sensitize the tumors to checkpoint inhibitors.

And one study we have initiated in Germany, the ADVANCE study, where we combine derazantinib with a checkpoint inhibitor, atezolizumab. And this might also be interesting because derazantinib has also activity against CSF1-R, which actually inhibits macrophages, and as you can see here in the multi-omic profiling these tumors are enriched in M2 macrophages. So maybe we have here 3 targets, macrophages, the driver, FGFR2, and also the checkpoint PD-L1. And the study is still ongoing.

As already mentioned, the KEYNOTE-966 study, which will most likely read out this year, so we are very much looking forward to this really large study. And IMbrave151 has been reported today. I think an interesting signal for overall survival, maybe, still not major. And if you’d like there are 2 experimental arms, so we need to pay attention for more major overall survival data. There are also combinations with local therapies, as we have just discussed, and what we can consider for patients with limited disease.

And just to take home, we have now a positive study, so I think durvalumab is now clearly part of the first-line standard. Gemcitabine and cisplatin we do not have good biomarkers, and I think we need to wait for the studies that will be reported in the near future. KEYNOTE-966 I think is a very good validation of IO in BTC, and other ongoing studies with locoregional therapies or targeted therapies I think are very interesting, but we need to wait to see more data. That’s it.

DR KELLEY: Thanks so much, Arndt. That was a great talk. I want to take 1 minute to ask a quick audience response question — or an audience question rather, because I think it’s a pretty common one. So rechallenge with gem/cis — so the TOPAZ trial, if everyone recalls, patients were randomized to gem/cis plus durva versus gem/cis, and then in patients treated with gem/cis/durva the treatment is the triplet for 6 months, or 8 cycles, and then maintenance durva alone.

So the question from the audience, which is a great one, is after say 6 months on maintenance durvalumab in a patient who previously had done quite well, who now progresses, do you go to a second-line FOLFOX or do you go back to gem/cis?

PROF VOGEL: So I would definitely go back to gem/cis. So I think we should really use our drugs as long as they are active, and if you do not use them in the maintenance phase they cannot work, right? So I would rechallenge, and I would not go to FOLFOX. Then in second line I mean irinotecan based or FOLFOX is —

DR KELLEY: Yeah, and of course it ends with the targeted therapy, which is a great lead-in to our next one.

Cases: A woman in her late 70s with intrahepatic cholangiocarcinoma considered unresectable — an FGFR2-KIAA1598 fusion was detected on NGS — and a man in his early 60s with metastatic cholangiocarcinoma with severe muscle pain on pemigatinib — Dr Fonkoua and Joseph Martins, MD

DR KELLEY: With that I’ll introduce Module 4, integration of targeted therapy into the management of advanced HCC from Dr Goyal. And again, to be preceded by some very interested cases from our faculty colleagues in the community.

DR LOVE: The next 2 cases are first, from Dr Fonkoua who is treating a 77-year-old woman with intrahepatic cholangiocarcinoma with an FGFR2 fusion. The patient was initially considered unresectable.

DR FONKOUA: The patient was 77, didn’t want to be too aggressive. She wanted to kind of preserve her quality of life. So we discussed options, gem/cis and pemigatinib.

She was very reluctant to try chemotherapy, so she favored targeted therapy, which we started initially at the standard higher dose of 13.5 mg daily, which she tolerated okay.

Some mild hyperphosphatemia, which we managed with dietary changes. And otherwise, we decreased the dose to 9 mg, and that worked out okay for her. It was better tolerated.

Clinically, she felt better, and then we got a PET after her 6-month restaging, and actually the avidity had resolved.

So we had her evaluated by a surgeon.

And she was taken for a left hepatectomy and then has been under surveillance.

DR FONKOUA: Is there any role of using cfDNA (free circulating tumor DNA) as part of the surveillance strategy, but also to pick up these potential clonal mutations that may evolve.

On the left here you can see — those are the baseline images, you can see kind of a regional mass.

And you can see the avidity.

After about 6 months of treatment, and you can see what’s left there. There’s some dystrophic calcification, and more importantly the avidity is almost completely gone.

So very impressive response with just pemigatinib.

DR LOVE: Dr Joseph Martins is managing a 60-year-old man with metastatic cholangiocarcinoma who’s also responding well to pemigatinib however the patient is having severe muscle pain.

DR MARTINS: I would have almost discussed hospice with him, but the next-gen sequencing and he had FGFR positive.

So I said, well let's give it a try. And it was literally probably just approved right when he was found to have the mutation. So he was obviously my first patient on that medicine. And immediately his tumor marker started coming down, his liver pain got better, and he's pretty much doing great. Now, cancer-wise he's doing great. He's having a lot of bone pain and muscle pain which is a listed side effect of that drug.

Have they seen muscle aches and muscle pain bad enough to cause a need for hydrocodone? And anything they can tell us about the other side effects. Like I have not had to deal with hyperphosphatemia but how do they deal with that?

DR KELLEY: Great. So we have 2 excellent cases pertaining to FGFR2 inhibition and thankfully experts here to help us with them. And so I think taking Dr Fonkoua’s case first the question pertains to surveillance in the adjuvant setting for a patient that was never expected to get to adjuvant but had a great response to pemigatinib. So I’ll turn to Dr Goyal to ask how would you respond to this question.

DR GOYAL: Yeah. Dr Fonkoua raises 2 very good questions, 1 is about minimal residual disease and then 1 is about evolution of clonal mutations, which one could consider around resistance. So certainly we know from the colon cancer literature that we’re using ctDNA to look for minimal residual disease and to monitor and to consider who to give adjuvant therapy to or not.

In someone who gets neoadjuvant pemigatinib or another FGFR inhibitor what is the role of adjuvant therapy in that setting? Would you restart pemigatinib? Would you do 6 months of capecitabine? It’s certainly a data-free zone because neoadjuvant therapy is not standard in cholangiocarcinoma, but I commend the creativity of Dr Fonkoua to be able to get this patient certainly to resection to be able to have her have curative-intent surgery.

Because it’s not standard it’s hard to know what to do with ctDNA results in the biliary tract cancer setting. To be very honest, I typically do not order ctDNA to use for surveillance after resection of biliary tract cancers, but it’s not unreasonable if the cfDNA is positive to consider either restarting pemigatinib, considering this patient does not want to do chemotherapy, or to consider doing chemotherapy in that case.

I think in terms of using cfDNA for monitoring of clonal mutations we usually see those in the setting of a patient being on the FGFR inhibitor, having the selective pressure on the rumor, and then developing resistance mutations in the kinase domain. In the postoperative setting it the patient is not on a FGFR inhibitor wouldn’t expect for that patient to develop new mutations in the kinase domain.

DR KELLEY: Great, thanks. I think for the second question about muscle aches and muscle pain I’ll direct that to Arndt, who has a lot of experience with pemigatinib from the Phase II trial and also clinically. So the question is have you seen significant treatment-limited muscle aches with pemigatinib requiring narcotics?

PROF VOGEL: So actually I have never seen it before. So of course we do have side effects. We have the stomatitis and nail toxicity, but — so this severe muscle pain I have not seen. Of course it can happen, obviously. In the end you need to see whether it can be treated by best supportive care if you’d like. Otherwise you need to dose reduce or maybe stop treatment.

DR KELLEY: Yeah, I agree. I haven’t seen it as well. Lipi?

DR GOYAL: I agree. I also have not seen it. The only thing I have seen is rhabdomyolysis once on an FGFR inhibitor, so maybe it would be worth checking a CK, and in 1 patient on pemigatinib after more than a year of therapy she started developing some arthritis in her joints, but not to the level of needing narcotics. So it might be worth like even having a rheumatology consult just to see if there isn’t some other diagnosis, given it’s considerably rare on pemigatinib.

DR KELLEY: That’s a great call.

Cases: A man in his early 50s with recurrent metastatic cholangiocarcinoma with a BRAF V600E mutation receives dabrafenib/trametinib and a woman in her mid 60s with cholangiocarcinoma with an IDH1 mutation receives ivosidenib — Farshid Dayyani, MD, PhD and Niyati A Nathwani, MD

DR KELLEY: All right. That brings us to our next cases.

DR LOVE: Dr Farshid Dayyani has a 53-year-old man with metastatic cholangiocarcinoma who has a BRAF V600E tumor mutation and is responding to dabrafenib/trametinib.

DR DAYYANI: We were able to get him started.

Dabrafenib trametinib, and he had a great quick response within 4 weeks CA19-9 normalized, stayed at 5, MRD assay turned negative. And he trucked along for about a year, 12 months.

Now more recently his Signatera has slowly gone up.

So I repeated his tumor agnostic liquid biopsy to see what's going on.

The BRAF clone, as you would expect, is gone in the repeat liquid biopsy and his original p53 mutant clone is taking over. So patient is doing well. He is still on dabrafenib/trametinib. His MRD is positive on a very low level, but positive is positive so sooner or later it will come.

The question is, do you switch anything or do you keep him on the current regimen?

DR LOVE: Dr Niyeti Nathwani has a 64-year-old woman with IDH1 mutated cholangiocarcinoma who is receiving the IDH1 inhibitor, ivosidenib.

DR NATHWANI: She’s currently getting concurrent treatment with radiation and capecitabine. We also did biomarker testing which did show this time the IDH1 mutation that was positive. And now that we have a targeted agent for that in this use, particularly in cholangiocarcinoma, ivosidenib.

Can we use this as like a therapy after completing concurrent chemoradiation was the question. I have never used any of these agents, so it was interesting that it’s from AML and then straight to cholangiocarcinoma, and of course side effects. And I’m guessing there’s no differentiation syndrome, but to prevent that are we going to use steroids or what their experience is with this.

DR KELLEY: All right. So we have 2 great questions and 2 very interesting targeted therapy cases. So I think starting with the BRAF inhibitor patient — BRAF mutation patient on BRAF/MEK inhibition with dabrafenib/trametinib from Dr Dayyani, and the question is in a patient who’s clinically doing well but has biomarker going up, in this case cell-free DNA. Some progression of a p53 mutation on cell-free DNA, as well, but otherwise feeling well with no symptoms of progression. I’ll direct this to Rich. Would you change therapy, or would you go with scans and clinical practice presentation?

DR FINN: Yeah. I don’t think we’re at the point where we would change things. And it’s the same thing with CA 19-9 now, right? Tumor markers make patients nervous, but I remind them they don’t get sick, they don’t get symptomatic from cell-free DNA, from an elevated CEA, or CA 19-9. They get sick and in trouble when we have objective progression.

I assume that at some point this data will mature and maybe indicate that there’s something else to do, but we’re not doing them any favor by telling them there’s smoke, and we’re just waiting for a fire. I mean I would sit tight. You’re feeling well, tumor’s controlled, continue current course.

DR KELLEY: I’ll add that I don’t monitor cell-free DNA absolute values as an index of treatment response in biliary tract cancer patients because I don’t think we have data that they are — just like PET scan FDG avidity we don’t have great data that they are correlated with radiographic or clinical response either.

So I guess I’ll ask Dr Goyal — Lipi and Arndt, any other thoughts on this?

DR GOYAL: Yeah. CtDNA is dynamic, and so you can check it on Monday and you might not see the BRAF clone and check it on Tuesday and you might see the BRAF clone. And so I totally agree with Rich and Katie that I would continue the dabrafenib and trametinib given clinically the patient’s doing so well. And also TP53 mutations, we often see them as subclonal mutations that emerge at the time of — when people have metastases. And again, ctDNA is so dynamic that it’s always hard to know when we order tests, always with our patients, like how do we convey those results and use those results in clinical decision making. But I would have — the patient’s symptoms and scans trump ctDNA in this case.

PROF VOGEL: I mean I completely agree, but maybe 1 point, I think it’s good that we try to understand resistance to these targeted therapies. So I would not monitor it during treatment. I mean that is smoke. I think it’s a good example. But once they have real progressive disease, yeah, I think it’s really a very good idea to perform either a liquid biopsy.

I mean we tend to biopsy the tumor, the progressive nodules, to really understand and correlate to the liquid biopsy, and I think that’s really important to understand the resistance mechanism and to see whether we can have additional targeted therapies. I mean for — just yesterday or 2 days ago published futibatinib in the New England Journal, and I think we do now have more and more evidence that we can sequentially use these targeted therapies, and for that I think it’s very good to redo NGS in patients on treatment.

DR KELLEY: Now the second case we were asked about is an IDH1 inhibitor from Dr Nathwani, and I think — I don’t think we have the full story on this patient. It’s a patient receiving concurrent radiation and capecitabine presumably for either adjuvant or definitive local therapy intent. So I guess first let’s assume this is a patient who’s got some unresectable disease and is being treated definitively with chemoradiation but also has an IDH1 mutation.

I think I’ll start with Lipika again. Would you add — after the patient has received definitive chemoradiation, either for local unresectable disease or adjuvant, would you add ivosidenib now or would you wait?

DR GOYAL: I think it’s a great creative question, and I think, as Arndt was saying to a previous question, it’s great that we have options. Now that we have an FDA-approved IDH1 inhibitor it’s great to consider it. I feel this question is really about maintenance; in someone who doesn’t have active disease is there a role for maintenance ivosidenib to prevent the emergence of progression. And I would say it’s a data-free zone.

The good news about ivosidenib is that it’s a very well-tolerated drug and therefore there was able to be a randomized trial against placebo where patients didn’t know which drug they were on, placebo versus IDH1 inhibitor ivosidenib. So I think it’s shared decision making with the patient.

The question though is if you start ivosidenib now are you going to get a longer overall survival compared to if you start ivosidenib at progression. Are you going to get a longer overall survival? We don’t know the answer to that, so I think either way, to be very honest, is very reasonable with decision making with the patient.

DR KELLEY: Great. Thanks so much. And I’ll add that we typically do not see differentiation syndrome in solid tumors treated with ivosidenib. That’s really restricted to hematologic.

So with that our last talk from Dr Goyal, from Lipika, which is Module 4, targeted therapy integration.

Faculty presentation: Dr Goyal

DR GOYAL: Thank you very much to Research To Practice for inviting me to give this talk. I’m really excited to talk about targeted therapy because targeted therapy has really transformed the way we think about biliary tract cancers. And people use the word transform all the time in oncology, and I will say I don’t think it has transformed outcomes for patients as a population of patients with biliary tract cancers, but I do think it has transformed the way we think about the disease. I think targeted therapy has transformed patient outcomes in non-small cell lung cancer in the way immunotherapy has transformed outcomes in melanoma and MSI-high positive tumors.

I think we have made progress, and we’ve improved outcomes for patients, but I think progress is like smoke before the fire. I think the transformation is coming because I think we’ve been able to find targets, that we just need to understand the science better and develop better drugs for, but I think the key is that we’re thinking about this disease differently and to now be talking about the progress we’ve made.

So I think we’ve heard about a couple of these today from Arndt. We’ve heard about gem/cis plus or minus durva in the chemotherapy bucket. We’ve heard about immunotherapy, obviously, from Arndt as well. And today I’ll be focused on targeted therapy.

So we’ve also heard from Arndt that there are multiple different anatomic subtypes of biliary tract cancer. In intrahepatic cholangiocarcinoma approximately 40-50% of patients have actionable alterations. In perihilar tumors and gallbladder cancer about 15-20% of patients have actionable alterations, and a little bit less in periampullary tumors.

What are those alterations? In intrahepatic cholangiocarcinoma we see IDH1 mutations and FGFR2 fusions as the 2 predominant alterations. And these are both therapeutically relevant and diagnostic because rarely do you see IDH1 mutations and FGFR2 fusions in other cancers. So if you have a solid mass in the liver it’s most likely intrahepatic cholangio.

We also see BRAF mutations across all 3 of these different subtypes, and the 1 alteration that we see more in extrahepatic cholangio and gallbladder cancer is HER2 amplification, and we’ll talk about the drugs for these different targets.

Because profiling is so important the Cholangiocarcinoma Foundation has started a campaign to educate patients about the importance of understanding their tumor and really advocating with their oncologist in order to get molecular profiling. We know even with ALK and EGFR inhibitors sometimes patients with non-small cell lung cancer are not undergoing molecular profiling. I think the same curve is going to happen with cholangiocarcinoma, where we have all these multiple positive trials. It’s going to take some time for all patients to get profiled who could be candidates.

Liquid biopsy has come up several times. The main way I use liquid biopsy is 1) the way Arndt was talking about, to look at resistance to targeted therapy, but then also in patients who get biopsies and the tissue profiling fails because cholangiocarcinoma, there’s a lot of necrosis and up to 25% of patients have failed profiling on tissue. This is a study of around 1,600 patients who got the liquid biopsy by Guardant, and 44% of patients had actionable alterations, in patients with biliary tract cancers.

It's important to know your assay because some mutations you can capture well, other mutations you can’t capture well. IDH1 mutations in this study, there was an 87% concordance between blood and tissue, but for FGFR2 fusions it was 18%. And so this is certainly improving as the technology for liquid biopsy improves, but it’s important to know that if you don’t see it on liquid biopsy it doesn’t mean it’s not there.

So there are different ways to study targeted therapy in uncommon cancers. As we know there are about 8 to 10,000 cases of biliary tract cancer in the US each year compared to over 100,000 cases of colorectal cancer and compared to over 200,000 cases of lung cancer.

So when you have a less common disease, and then you have specific targets, how do you study this? So when you have a high enough frequency, like FGFR2 fusions and IDH1 mutations, you can have target-specific trials, and I’ll be starting with that. And then for rare targets you can have basket trials, and then in the UK there is an umbrella trial where all patients with biliary tract cancer can enter, and then depending on their mutation can go into one of several arms.

So I’ll start briefly with FGFR and IDH1. I will say that I have about 20 trials to discuss, so with each trilaciclib I will talk very superficially about the trial and mainly the efficacy data; will not be talking about any safety data. And you’ll have all of these slides, so you can come back to the waterfall plots and the data afterwards. But it’s more just to give you the breadth of the number of drugs that have been FDA approved and are on the NCCN Guidelines.

So up until 2020 — or up until 2020 we didn’t have any drugs approved in cholangiocarcinoma. The first drug ever to get approved was pemigatinib, and close behind it was infigratinib. These are both irreversible inhibitors — sorry, reversible FGFR inhibitors that target FGFR2 fusions and rearrangements. The response rates were 36, 37% for pemigatinib and 23% for infigratinib. Pemigatinib is still commercially — is commercially available, and infigratinib the team has pulled commercial development of that drug.  

Just yesterday the results of futibatinib were published from the FOENIX-CCA2 study. This is an irreversible FGFR inhibitor, and the overall response rate was 42%. So we now have 2 FDA-approved options for FGFR2 fusion and rearrangement positive cholangiocarcinoma. It’s 1 target that if I see this on NGS it’s the kind of thing I will call my patient on a weekend because I’m quite excited that they have this alteration because it opens up lots of therapy options.

There are multiple new FGFR inhibitors that are being developed. By the team from Relay there’s the FGFR2-specific inhibitor, RLY-4008, and in this study patients who got the RP2D of 70 mg there was an overall response rate of 82%. This is preliminary data. This was in 17 patients. This study is ongoing, and we look forward to the mature results, including the duration of response.

This is data that was just presented here at GI ASCO on a drug called tinengotinib, a small molecule inhibitor that’s a reversible inhibitor, and in patients who previously got FGFR inhibitors were seeing some responses. And so this drug is continuing to be developed in FGFR2 fusion-positive cholangiocarcinoma and in patients with other FGFR alterations. But it’s nice, as Arndt was saying, to be able to sequence FGFR inhibitors, so it’s good to have options like this.

These are data also presented at GI ASCO today; erdafitinib. The response rate in FGFR2 fusion-positive cholangiocarcinoma. And 91% of patients had FGFR2 fusions, 9% of patients had mutations in either FGFR2 or 3. And in this study the response rate was 60%, duration of response 5.5 months.

Ivosidenib. This is the one study in targeted therapy that was a randomized Phase III trial, 2:1 randomization of the IDH1 inhibitor versus placebo. And in this the primary endpoint was progression-free survival, and it met its primary endpoint with ivosidenib showing a PFS of 2.7 months versus placebo 1.4 months. And this is FDA approved.

And then for the more rare — the more rare targets. We’ll talk about these. BRAF V600E about 4-5% of biliary tract cancers. Patients in this basket study got dabrafenib plus trametinib. The response rate was around 50%, and this has gotten a disease agnostic approval by the FDA.

For HER2 overexpression and amplification there are 3 different drugs or drug combinations that are in development. Trastuzumab/pertuzumab has gotten onto the NCCN Guidelines. You can see the overall response rates with these 3 regimens are between 20 and 40%. And trastuzumab deruxtecan and zanidatamab are both in development in biliary tract cancer.

And then NTRK and RET fusions, very uncommon, less than 1%, but really important that we do look for them because people have remarkable responses, over 50% in these disease agnostic basket studies. And this also is FDA approved in a disease agnostic manner.

There are many other alterations. I know we mentioned PIK3CA in HCC, and it’s true that PIK3CA mutations only have approval for therapy in breast cancer. I think one of the important things about putting people on clinical trials is you can do discovery because certainly in cholangiocarcinoma I’ve had patients who’ve gotten 8 months of PFS or 6 months of PFS on PI3K inhibitors, so every once in a while you can get lucky. So I would say in the context of a clinical trial it’s really helpful to look at these.

So these are the NCCN Guidelines. We’re really happy that it’s getting more and more crowded over time because this used to be quite a white slide.

And my final slide is that this is what we had for about 10 years in cholangiocarcinoma; we had chemotherapy. And as you can see the space is getting more crowded now with many different options. So we’re certainly making progress. We would like to see the kind of progress where we hit homeruns that Dr Finn was talking about in cholangiocarcinoma — in HCC, but overall the most important thing is that we profile all of our patients because that’s how we find targets where we can help people and consider patients for a clinical trial so we can get people — get drugs to FDA approval more quickly.

Thank you very much.

DR KELLEY: Thanks, Lipika, that was a great talk, and it’s amazingly right on time, 7:30, after the end of a long week, and I know many of you have been traveling. So we will conclude tonight’s session with a huge thank you to all of you here in the live audience, as well as all of you here in the Zoom audience, whatever time zone you’re in.

Thank you very much for your attention. Good night.