JCO PO author Dr. Mark Yarchoan, MD, of Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, shares research on the four distinct subsets of biliary tract cancers (BTCs) and their varying immune responses and mutations. Dr. Naqash and Dr. Yarchoan discuss the article’s study of intrahepatic cholangiocarcinoma [IHC], extrahepatic cholangiocarcinoma, and gallbladder cancer. Click here to read the article!

Dr. Naqash: Welcome to ASCO's JCO Precision Oncology Conversations where we bring you the highlights and overview of precision oncology. Episodes will feature engaging conversations with authors of clinically relevant and highly impactful and significant JCO PO articles.

These articles can be accessed at ascopubs.org/journals/po. Hi, I am Dr. Rafeh Naqash, Assistant Professor of Medicine at Oklahoma University Stephenson Cancer Center, where I focus on Phase I clinical trials and lung cancer. You're listening to JCO Precision Oncology Conversations podcast.

Today, I will be talking with Dr. Mark Yarchoan about his recent paper, Clinical, Genomic, and Transcriptomic Data Profiling of Biliary Tract Cancer Reveals Subtype-Specific Immune Signatures.

Dr. Yarchoan is an Associate Professor of Medicine in the GI Medical Oncology branch at the Johns Hopkins, Sydney Kimmel Comprehensive Cancer Center.

Full disclosure for our guests will be linked in the show notes and can be found on the article's publication page. Dr. Yarchoan, thank you so much for joining us today.

Dr. Yarchoan: Thank you for having me.

Dr. Naqash: For the sake of our listeners, Mark, could you tell us about your professional background and your research interests?

Dr. Yarchoan: Sure. I'm a medical oncologist by training focused on hepatobiliary malignancies at the Johns Hopkins Hospital, a co-director of Liver Multidisciplinary Clinic, and I'm interested in novel immunotherapeutic approaches to treating liver cancers.

Dr. Naqash: That is definitely an unmet need, which we're going to talk about more as we discuss your paper.

Again, for the sake of our listeners, before we take a deeper dive into the published paper, could you provide us a one-minute overview of the main findings from your manuscript that was published in JCO Precision Oncology?

Dr. Yarchoan: Absolutely. So, we looked at the molecular subtypes of cholangiocarcinoma. So, there are essentially two common forms of liver cancer. There's hepatocellular carcinoma and biliary tract cancers, also sometimes called cholangiocarcinomas.

Cholangiocarcinomas are the rarer of the two types of liver cancer and we know that cholangiocarcinoma’s not really one cancer, it's a number of different cancers that all anatomically start within the bile ducts of the liver.

And what we did working with a number of different collaborators and then also, with Tempus, which is a major molecular profiling company, is we looked at the landscape of mutations that occur in cholangiocarcinoma, and then tried to correlate these mutations with features of the tumor immune microenvironment using RNA data from the same tumors.

Dr. Naqash: Thank you so much. So, again, being a thoracic oncologist myself, we have a lot of excitement associated with immune checkpoint alone or immune checkpoint in combination with chemotherapy, similar to what has been seen in some of the upper GI cancers in the last couple of years.

Now, before you started this paper or this work with Tempus, what was known about the landscape of biliary tract cancers and what have been the standard approaches of management for advanced biliary tract cancers?

Dr. Yarchoan: You know, many of us who treat biliary tract cancers have joked for years that we treat the lung cancer of the GI cancer world because so many of these cancers have actionable molecular alterations.

I think that's something that has been known for a long time. Over the last couple of years, for the first time, we have multiple drug approvals in cholangiocarcinoma for molecularly defined subsets of disease.

So, for example, we have now inhibitors of FGFR2 that have been approved for FGFR2 fusions or rearrangements. We have an IDH1 inhibitor approved for cholangiocarcinoma.

There are some tumor agnostic approvals like for BRAF and NTRK, all of which we do occasionally find in biliary tract cancer.

So, you know I think it's been known for a long time that looking for molecular alterations is fruitful in cholangiocarcinoma. You often find these things and these cancers often respond to these agents.

I think what is unique here is we've taken this knowledge and first of all, this is the largest series of biliary tract cancer that has ever been profiled to our knowledge, and other profiling efforts have been international.

But we know that biliary tract cancer in Southeast Asia can sometimes be different from what we see here in the U.S. because there, those cancers are often caused by liver flukes, which we don't tend to see as much in the U.S. So, the genetics can be somewhat different.

And finally, I think what is unique about our paper is that we have actually correlated these molecular differences, these molecular drivers of cholangiocarcinoma.

And we've correlated these drivers with the RNA from the same tumor to really understand when you have a tumor that is for example, FGFR2 fusion or rearrangement, what is typically found in the immune microenvironment from that same tumor.

Dr. Naqash: Thank you for that explanation. Now, going back to why biliary tract cancer’s potentially not as responsive to immune checkpoint inhibitors as some of the other GI-based tumors — pancreatic cancer, as you very well know, has a certain tumor microenvironment associated with presence of fibroblasts and other immunosuppressive macrophages that result in lack of response to immunotherapy-based combinations.

What is unique in biliary tract cancers, as you've nicely explained in the discussion or your introduction, that based on Keynote-158, the response rates have generally been in the range of 2 to 10%. So, what is unique here that results in tumors being not so responsive to immunotherapy-based combinations?

Dr. Yarchoan: Yeah, biliary tract cancers tend to be fairly immune resistant cancers. As you mentioned, the response rate to single agent anti PD-1 immunotherapy is probably in a range of about 5 to 10% in most studies.

We do have new data now from the TOPAZ-1 study, this is a randomized first line study of Gemcitabine Platinum with or without durvalumab, which is a PDL1 inhibitor where the addition of durvalumab did modestly improve survival.

So, there is some activity for immunotherapy, but it's quite low as you mentioned. And again, I think that we're still learning why this is, but cholangiocarcinoma tends to have, number one, a very low tumor mutation burden.

So, the total number of mutations, which is a surrogate for the number of neoantigens that the immune system can see tends to be low. But also, when we look at the immune microenvironment at these cancers, we tend to find relatively low numbers of effector T cells in the microenvironment, which are an anti-tumor subset.

We tend to find high numbers of immunosuppressive cell types, M2 macrophages and T regulatory cells. And these tumors also tend to have a lot of stroma which can impede the immune system coming in.

So, the two major accesses that can determine immune sensitivity, the mutational burden in the immune cells, and the microenvironment, both point to an immune suppressed tumor type.

Dr. Naqash: Totally agree. And I think that's why findings that you describe in this paper have a lot of relevance to precision medicine and how we can approach biliary tract cancers, not just anatomically but also from a molecular standpoint.

Going back to lung cancer analogy, small cell lung cancer until a few years back was considered one entity and now, it is three to four different entities based on some transcriptional factors which determine how patients respond. And I believe there is something similar going on here like you described.

So, now taking a deeper dive into the paper, you did mutational signature analysis clustering, and then you also assessed RNA. What would some of the mutation-specific findings associated with whether tumors were intrahepatic or extrahepatic gallbladder in this manuscript?

Dr. Yarchoan: Sure. So, we, as you mentioned, did some cluster analysis looking at the driver mutations in cholangiocarcinoma. And what we found were essentially, there appeared to be four distinct subsets of these cancers.

One subset which has been reported in other similar cluster analyses, including an international cluster analysis; one subset really appears to be driven by FGFR2 fusions or rearrangements and is entirely an intrahepatic cholangiocarcinoma subset.

And we found that this subset tended to have the lowest immune infiltration of any of the four subsets that we looked at.

A second subset, which we called mutational cluster three, was both intrahepatic and extrahepatic disease. And the most common mutation was really in CDKN2A or N2B.

The second cluster that we looked at, clusters 1, 2, 3, 4 — I think is one of the most interesting clusters. And this is a cluster that was enriched for genes that seemed to alter the epigenome of the tumor.

So, for example, IDH1 mutations, which we know result in an oncometabolite 2HG were common in the subset also ARID1A and PBRM1 were common.

And then finally, the largest subset of all were the tumors that had frequent mutations in KRAS, TP53 among other mutations. We found that this was the most inflamed cluster, had the highest PD-L1 expression, slightly enriched in the extrahepatic subset versus intrahepatic subset.

Dr. Naqash: Definitely interesting findings based on the cluster analysis and the mutations in each cluster.

Going to cluster one where you have found that TP53, ATM, et cetera, mutations are commonly altered genes, are you suggesting based on some of this work that the immune inflamed tumor microenvironment could potentially be linked to TP53, which is cell cycle checkpoint, which when altered could result in higher DNA damage similarly with ATM being in the DNA damage pathway.

Could that be a potential explanation for why these alterations are resulting in a hot tumor immune microenvironment?

Dr. Yarchoan: Certainly, could be. You know, obviously, we can't prove that. This was correlative, but I think that would certainly explain what we found.

Dr. Naqash: And again, going back to cluster four with FGFR2 and BAP1 alterations, which had the lowest PD-L1 and lowest immune microenvironment inflammatory signature, is that potentially a cluster that you think would most commonly benefit from targeted therapies rather than immune checkpoint-based therapies?

Dr. Yarchoan: That's our hypothesis here. This was a subgroup that appeared to be non-inflamed, had a low mutational burden. Genomically, I would call this a stupid cancer in the sense that there's a very powerful driver mutation and not a whole lot else going on here.

It seems like the sort of tumor that may benefit from targeted therapy. I think what we don't know, and one of the questions that immediately comes out of this work is, is FGFR2 itself immunosuppressive and that's why we don't find a lot of immune cells.

And if so, it raises the question whether these sort of cluster four tumors would benefit from a combination approach of targeted therapy plus immunotherapy. And so, I think that's an open question, something that we hope to look into.

Dr. Naqash: Definitely, leads to potential subsequent work in this field to expand more on these findings.

Now, going back to some of the other important findings in your paper, you described actionable biomarkers in the biliary tract cancers based on proven trials with specific therapies.

One of the questions that I had was you take TMB ≥10 as a potentially actionable biomarker. Is that true for biliary tract cancers? Because there's data from different groups showing that different TMB cutoffs have different meaning for different types of tumors.

So, is 10 a cutoff where you are reasonably comfortable in saying that potentially, single agent immunotherapy may work and TMB here is predictive of outcomes, or do you think a higher TMB is probably more relevant in these cancers that are generally not immune responsive?

Dr. Yarchoan: Yeah, the approval for pembrolizumab in high TMB tumors has been controversial. I think probably more controversial than it should be. But Keynote-158, which was sort of a prospectively analyzed study of pembrolizumab and multiple tumor histologies, actually included biliary tract cancers.

And the response rate in that study for biliary tract cancer in the TMB high group again, small numbers of patients but, was consistent with pembrolizumab being very active.

I have to say in my own clinic, I've had some absolutely spectacular responses for high TMB biliary tract cancer, and so I tend to believe the data. I've actually had complete responses. I've had patients downstage to surgery who went on single agent anti PD-1.

So, I think this is an important group of patients and I think it is a group that we shouldn't miss for biliary tract cancer.

Dr. Naqash: Definitely agree with you. All of us strive as medical oncologists to get some of those complete responses that don't come often very commonly or easily. But it's always nice to see those responses and individuals especially with these biomarker-specific tumors with high TMB or high PD-L1.

Now, other findings from your paper, you describe mutual exclusivity of two mutations in a certain subgroup, which was TP53 and BAP1. Could you explain the clinical relevance or the molecular significance of these mutually exclusive mutations and what they might mean in the context of this tumor?

Dr. Yarchoan:The short answer is these were clearly in different immune subsets and different molecular subsets. Again, the TP53 belonged to what we called cluster one. Usually, went with KRAS, was more common in extrahepatic disease.

The BAP1 mutations tended to go along with FGFR2 fusions or rearrangements, was intrahepatic. But exactly why these were mutually exclusive, why was it not evolutionarily advantageous for tumors to have both of these things at once?

You know, I think remains an unanswered question. There’s certainly a hypothesis for why the tumor wouldn't need both, but I don't know.

Dr. Naqash: And you also looked at PD-L1 and tumor mutational burden across the biliary tract tumor subtypes. Could you tell us more about some of those findings?

Dr. Yarchoan: Yeah, for sure. I mean, I think our cluster one, which again was the most inflamed cluster, had more KRAS, more TP53 — not surprisingly seemed to have the most immune infiltration, the highest PD-L1 expression in a trend towards a higher tumor mutation burden as well. Not totally surprising.

I don't want to over-interpret that finding though because we've looked now at the prospective data from TOPAZ-1 where patients got GemCis with or without durvalumab. As a matter of fact, PD-L1 was not a very strong predictor for durvalumab benefit in that study.

So, I wouldn't over-interpret the trend towards higher PD-L1 expression in that subset. It may not be very clinically actionable.

Dr. Naqash: Sure. Now, another finding is the TMB comparisons across the biliary tract cancer subtypes. And to me, it seemed like the TMB, median TMB, was slightly different but not significantly different that may not necessarily have a huge clinical relevance. Do you agree with that? What are your thoughts on that side?

Dr. Yarchoan: We’re in an era now where we just get TMB at an individual patient level. So, I don't think we're going to be relying on cluster analysis to figure out a patient's TMB.

So, I don't think it's necessarily clinically actionable. I do think there were some findings here that are interesting from a research perspective.

For example, ARID1A was associated with a higher tumor mutation burden; something that's been reported in other tumor types and that's certainly interesting and something that should be followed up on, I think, from a research standpoint.

Dr. Naqash: Excellent. Now, last figure that I came across, which was again, very interesting, was that you analyze association between all the different mutations and different immune gene signatures.

And it seemed based on the figure that TP53, again, stood out when you correlated with PD-L1 or CDA T cell or other immune inflammatory gene signatures.

Is there something going on with TP53 here that you think is potentially relevant co-mutation in this tumor, which could lead to some novel therapeutic combination approaches?

Dr. Yarchoan: Yeah, I don't want to over-interpret our work, but I agree, it's very interesting and this is a topic that is being investigated in many tumor types because TP53 is something that appears in many tumors including the ones that you treat.

So, I think this is particularly relevant because we have a bunch of TP53 targeted therapies coming down the pike. How those will modulate the tumor immune microenvironment is an interesting topic.

So, maybe just for sake of discussion, are you finding the same thing in lung cancer and what's the thought process in lung cancer now?

Dr. Naqash: Yes, again, very interesting question. So, we actually had some data last year that we presented in the context of lung cancer in STK11, which is again, commonly found in pancreatic and biliary cancers also.

And we saw TP53 co-mutated tumors had a very high immune inflammatory gene signature with all the different aspects of immune infiltration or Interferon gamma upregulation (IFN-γ).

So, likely, our assessment was that this was potentially linked to the cell cycle aspect and higher neoantigens for DNA damage. So, something common probably going on here in this tumor type as well as you mentioned.

So, one of the other things that I came across, which is interesting and I've come across a patient actually a few years back with cholangiocarcinoma, but that was associated with a liver fluke.

The tumors that you assess using the Tempus database obviously, were from North America and did not have Asian patients. And from my understanding, TOPAZ-1 did include a decent number of patients from Asia.

So, have you looked at or is there a potential reason to believe that liver fluke-associated tumors may have a higher inflammatory signature or has there been any data in comparing liver fluke versus non-liver fluke-associated biliary tract tumors to see what could be the genomic transcriptomic differences there?

Dr. Yarchoan: Yeah, as you mentioned, our study was purely for North American patients. It's not that there were no Asian patients. There certainly were Asian patients, including patients who may have been born in Southeast Asia but immigrated here.

There have been other analyses of the molecular landscape of biliary tract cancer that were international and included more patients from Asia, including a paper in Cancer Discovery a couple of years ago that I think is worth reading.

Overall, I think there is some data that this subgroup of patients may have a unique tumor immune microenvironment and unique mutational landscape. We tend to find more for two in these patients. They do tend to be, I think, a little more inflamed, have more PD-L1 expression.

Interestingly, in the TOPAZ-1 study, which again, was the prospective study of durvalumab, there appeared to be more benefit to durvalumab in patients in Asia than patients in the West. And it does raise the question whether that could be mediated in part by this liver fluke ideology.

So, again, certainly not a settled question, but something interesting that should be followed up on.

Dr. Naqash:Definitely interesting. I would again like to thank you Dr. Yarchoan and your team for this excellent work and sharing all your thoughts with us today, and giving us more insights into biliary tract cancers.

On behalf of JCO PO, I'd like to thank you for considering JCO PO as the final destination for this interesting work, and hopefully, you consider JCO PO for future work as well.

Dr. Yarchoan: Well, thank you for highlighting our work and for the opportunity to publish, and thanks for this discussion.

Dr. Naqash: Absolutely. It's been a pleasure talking to you today, and hopefully, our audience will find this conversation equally intriguing and interesting.

Thank you for listening to JCO Precision Oncology Conversations. You can find all our shows, including this one at asco.org/podcasts or wherever you get your podcasts.

To stay up to date, be sure to follow and share JCO PO content on Twitter. Our Twitter handle is @JCOPO_ASCO. All JCO PO articles and series can be found at ascopubs.org/journals/po.

Voiceover: The purpose of this podcast is to educate and to inform. This is not a substitute for professional medical care and is not intended for use in the diagnosis or treatment of individual conditions.

Guests on this podcast express their own opinions, experience, and conclusions. Guest statements on the podcast do not express the opinions of ASCO. The mention of any product, service, organization, activity, or therapy should not be construed as an ASCO endorsement.

Disclosures

Mark Yarchoan

Consulting or Advisory Role: Eisai, Exelixis, AstraZeneca, Genentech/Roche, Replimune, Hepion Pharmaceuticals

Research Funding: Bristol Myers Squibb (Inst), Merck (Inst), Exelixis (Inst), Incyte (Inst)

Guest Bio

Dr. Mark Yarchoan, MD, is a medical oncologist focused on hepatobiliary malignancies at the Johns Hopkins Hospital and co-director of Liver Multidisciplinary Clinic.

Clinical, Genomic, and Transcriptomic Data Profiling of Biliary Tract Cancer Reveals Subtype-Specific Immune Signatures with Dr. Mark Yarchoan

JCO Precision Oncology Conversations