JCO PO author Dr. Amit Mahipal shares insights into his JCO PO article, “Tumor Mutational Burden in Real-world Patients with Pancreatic Cancer: Genomic Alterations and Predictive Value for Immune Checkpoint Inhibitor Effectiveness.” Host Dr. Rafeh Naqash and Dr Mahipal discuss real world evidence of immune checkpoint inhibitors in pancreatic ductal adenocarcinoma.

TRANSCRIPT

Dr. Rafeh Naqash: Welcome to JCO Precision Oncology Conversations, where we bring you engaging conversations with authors of clinically relevant and highly significant JCO PO articles. I'm your host, Dr. Rafeh Naqash, Social Media Editor for JCO Precision Oncology and Assistant Professor at the OU Health Stephenson Cancer Center, University of Oklahoma. Today we are joined by Dr. Amit Mahipal, Professor of Medicine and Director of GI Oncology at the Case Western Reserve University in Seidman Cancer Center. Dr. Mahipal is also the author of the JCO Precision Oncology article titled "Tumor Mutational Burden in Real World Patients with Pancreatic Cancer: Genomic Alterations and Predictive Value of Immune Checkpoint Inhibitor Effectiveness." 

Our guest disclosures will be linked in the transcript. For the sake of this conversation, we will refer to each other using our first names. So Amit, welcome to our podcast and thank you for joining us today.

Dr. Amit Mahipal: Thanks for having me here.

Dr. Rafeh Naqash: Excellent. We came across your article in JCO Precision Oncology and it really aroused my interest because the topic and the audience that it caters to is very important in the current times. Because immunotherapy generally is considered- pancreas cancer the graveyard in immunotherapy in essence, based on what I have seen or what I have encountered. And now you're the expert here who sees people with pancreas cancer or has done a lot of work in pancreas cancer research side. So can you tell us the context of this work and why you wanted to look at immune checkpoint inhibitors in pancreas cancer?

Dr. Amit Mahipal: Absolutely, Rafeh. As you mentioned, pancreatic cancer is considered a what we call "cold tumors." They don’t typically respond to immunotherapy. And when we talk to our patients or patient advocates, as you know, patients are very excited about immunotherapy. Immunotherapy has transformed the treatment for a lot of different cancers and not only has increased survival, but the quality of life is so much different than with chemotherapy. This work came from based on the KEYNOTE-158 trial, which was a tumor-agnostic trial which accrued patients who had TMB high tumor. What that means is that tumor mutation had more than 10 mutations per megabase. And what happens is because of that trial, more than 200 patient trial, the FDA actually approved this immunotherapy or pembrolizumab as a single agent pembrolizumab for any patient with a solid tumor who has high TMB. Again, tumor mutation burden, more than 10 mut/Mb.

This question comes in now. Does this apply to our pancreatic cancer patient groups? Especially as we know these are "cold tumors" that typically do not respond. There have been multiple trials looking at immunotherapy, single agent, dual immunotherapy agents, as well as combinations with chemotherapy, with somewhat very, very limited success. So that was kind of the basis. So we wanted to look at this retrospective kind of review of a big database to see how many patients we can find who have high TMB and see in that patient population is immunotherapy really active based on the FDA approval or is pancreatic cancer not a tumor where we should try immunotherapy unit as a selective group.   

Dr. Rafeh Naqash: Thank you for that explanation. Taking a step back again, since you see these individuals with pancreatic cancer I imagine day in and day out in the space of drug development, what is the general current standard of care approach for individuals with pancreas cancer in your clinic? I'm talking about what are the most common approaches that you utilize that seem to be working or have FDA approvals in the pancreas cancer space.

Dr. Amit Mahipal: As with any tumor, the first thing is obviously staging. So depending on whether we're dealing with early stage or advanced stage and what are the goals of treatment. At this point, the only thing that can cure pancreatic cancer patients that would be considered conventional therapy is surgical resection. So any patient who is a candidate for surgical resection is in a different bucket compared to advanced patients. For early stage patients, we try to do what we call neoadjuvant treatment or neoadjuvant chemotherapy. We shrink the tumor or at least maintain it, look at the biology of the disease, and then take them to surgery, which typically involves a Whipple procedure if it's a head of the pancreatic mass.  

Moving on to advanced patients, that’s where we know the goal of treatment is palliative to increase survival, but unfortunately, most of the times we cannot cure them. And there the standard of care options include systemic chemotherapy. We have two typical regimens that we use, one is called FOLFIRINOX, which is a three-drug regimen of 5-fluorouracil, leucovorin, oxaliplatin, and irinotecan. And another regimen is gemcitabine plus abraxane, which is a two-drug regimen of gemcitabine plus abraxane. These are considered the standard of care. Unfortunately, the median survival even with the best standard of care chemotherapy is only about a year, 12-13 months, depending on what trials we look at. 

Dr. Rafeh Naqash: I still remember some of these regimens from my fellowship, where we had to decide which to give to each individual based on their performance status and clinical status, etc. But now I can see a lot of ongoing drug development in the space of pancreas cancer. I'm guessing that's why you wanted to assess both the molecular genomic landscape of pancreas cancer in this study and also look at the immune biomarker aspect. Could you tell us a little bit about the Foundation Medicine Clinical Genomic Database? How did you identify the patients, how many patients did you identify, what you narrowed down in the criteria, and the eventual sample size of what you were looking at? 

Dr. Amit Mahipal: FoundationOne has a rich database. They have two or three things. One is a genomic database only. So in our clinical practice, I think it’s some sort of next-generation sequencing or mutational testing for all patients with advanced solid tumors. All of these goes into their database. All of the samples that are sent to FoundationOne that goes into their database where they know the diagnosis of the patient and the know the sequencing results of these patients. In addition, they also have a clinical database called Flatiron. Basically, they collaborated with them. Flatiron has about 280 or so cancer clinics throughout the country, so a lot of community settings and some academic sites as well. They did not only have a genomic database, but they actually have a clinical database. They have demographics, clinical features, baseline clinical features, comorbidities, what kind of treatment they received, what would be the stage of the cancer, how many months of treatment they received, and their overall survival, and so on. So from that perspective, the FoundationOne has access to this partnership with Flatiron, clinical genomic database where they have both clinical data as well as genomic database for a lot of these patients. 

In our study, we only focused on patients with advanced pancreatic cancer. We excluded a lot of patients who did not have sequencing results available, they cannot be performed due to lack of tissue. So the first we talked about the genomic database and we found about about 21,932 patients, so almost 22,000 patients and there we had the sequencing and we also had the data on TMB or tumor mutational burden. So here, we classified them into two groups: high TMB and low TMB. High TMB was seen in 1.3% of the patients, and about 98.3% of the patients had low TMB. Here we looked at the genomic alterations between the two groups. So these are like our genomic group, so to speak of about 22,000 patients. And among them, as mentioned, that the clinical data was available for about 3300 patients or 3279 patients to be exact.  

After excluding some of those patients, we found about 51 patients who received immunotherapy. And when we say immunotherapy, it is single agent immuno checkpoint inhibitor like pembrolizumab or nivolumab. And then we classified them into high TMB versus low TMB and then we also looked at patients with high TMB and compared them to who received immunotherapy versus other therapies. Just to recap, we had about 22,000 patients where we have the genomic database and about 3300 or so patients who we have both genomic and clinical data for this patients. One of the key findings was that high TMB was present in only 1.3% of the patients, or about 293 patients out of 21,932. 

Dr. Rafeh Naqash: Definitely an interesting sample size that you had utilizing this resource, which, of course, is more or less real-world. It is important to gather real-world outcomes that you did.  

So, going to the TMB story of this paper, where you looked at immune checkpoint inhibitor use in these individuals, was there a reason why some of the individuals with low TMB were also given immune checkpoint inhibitors? From my understanding, I did see some checkpoint inhibitor use there. What could be the explanation for that?

Dr. Amit Mahipal: So this data is from 2014 to 2022. So from the span of about eight or so years. And as you know, immune checkpoint inhibitors were approved in the last decade. And there were a lot of not only trials, but even in the non-trial setting, people had tried immune checkpoint inhibitors in, frankly, different tumor types because of the success in some of the common tumor types, like melanomas, lung cancer, and so on. So I agree, as of today, we probably would not use immune checkpoint inhibitors in patients with low TMB or MSS. But at that time, I think that information was not available. So people with low TMB and MSI-stable tumors also received immune checkpoint inhibitors. But those numbers are again low. So it's not very high numbers.

Dr. Rafeh Naqash: Understandable. That makes it a little more clear. 

Now, you looked at the TMB aspect. I'm guessing you also looked at the MSI aspect of PDAC. What is your understanding, or what was your understanding before this study, and how did it enhance your understanding of the MSI aspect of PDAC? And I'm again guessing, since TMB high individuals are on the lower side percentage, so MSI high is likely to be low as well. Did you see any interaction between those MSI highs and the TMB highs on the PDAC side? 

Dr. Amit Mahipal: Yeah, absolutely. So we are very excited in general about MSI-high tumors for solid tumors because of their response to immunotherapy. Although I would do a caveat because we still don't know how MSI-high pancreatic cancer responds although there have been some real-world, very, very small series as well. In this study, one of the things is, is high TMB totally driven by MSI-high? That's a question that comes up, and TMB high may not matter. It's only the MSI-high that might matter. So definitely when we look at this patient population, we found that the patients who were 35-36% of patients who were TMB high also had MSI-high patients. So we do expect MSI-high patients to have a higher TMB compared to MSS patients. But there were about 66 or two-thirds of the patients who did not have MSI-high tumors and still had high TMB, as defined by, again, ten mutations per megabase. So we did see patients with MSI-stable tumors who had high TMB. And I think that was one of our biggest questions. I think MSI-high patients, we all tend to think that we would try immunotherapy even if it's in pancreatic cancer. I think what is not clear, at least from the real-world or any of the trial data, is if we were to give MSI-stable patients who have high TMB, if we give immunotherapy, are there any responses or any disease control that we see? And that was one of the reasons for this study. 

Dr. Rafeh Naqash: Now, one of the things that comes to mind, and again, I think you based it on the FDA approval for TMB high, which is ten mutations per megabase, as you defined earlier. I do a lot of biomarker research, and oftentimes you come across this aspect of binary versus a linear biomarker, in this case being TMB, where about ten, less than ten. Do you think, in general, an approach where you maybe have tertiles or quartiles or a biomarker, or perhaps a better approach in trying to stratify individuals who may or may not benefit from immunotherapy?

Dr. Amit Mahipal: That's a great point. I think when we use ten mutations per megabase as a biomarker, as a binary endpoint, do we apply it to all tumor types? I don't think that's a fair comparison, frankly speaking. We do know that high TMB, even in different tumor types, do tend to respond a little bit better to or do have better outcomes for patients treated with immune checkpoint inhibitors in different tumor types. But what that cutoff is not known in most of the tumor types. And also, one of the problems is how do you measure TMB and is it standard across different platforms? Like I'm just giving some names like FoundationOne, Tempus, Caris, and some obviously like MSKCC and some other university-owned panels as well. And frankly, I think if you look at different panels and if you send the same tumor tissue, you will get different measurements. So I think standardization is a problem as well. 

In one of the studies involving cholangiocarcinoma, for example, we found that a TMB of 5 was enough to have an additive effect of immunotherapy, same with chemotherapy, so to speak. But again, this needs to be validated. 

So you're absolutely correct. I don't know why we use the binary endpoint, but on the same token, the binary endpoint is easy to understand as a clinician. Like, “Hey, someone has this, do this, not this.” And when we look into a continuous range, I think the benefit obviously varies between high and low, different tertiles, and becomes somewhat challenging. How do you classify patients and what treatments to give? So I think in clinical decision-making, we like the cutoffs, but I think in reality, I don't know if the cutoff is a true representation. And maybe with the more use of AI or computing, we can just input some values, and then it can tell us what the best treatment option might be for the patient. But that's way in the future.

Dr. Rafeh Naqash: That would definitely be the futuristic approach of incorporating AI, machine learning perhaps, or even digital pathology slides in these individuals to ascertain which individuals benefit. 

Going back to your paper, could you highlight some of the most important results that you identified as far as which individual is better, whether it was immunotherapy, and you've also looked at some of the mutation co-mutation status. Could you highlight that for our listeners?

Dr. Amit Mahipal: So the first thing we looked at was the genomic database of almost 22,000 patients, and then we classified them into high TMB and low TMB, with about 300 patients in the high TMB group and the rest in the low TMB group. And what we found was, talking about again in the genomic database, that patients who have high TMB actually have low KRAS mutation. So if we think about KRAS mutation, pancreatic cancer, almost 85% or so of patients have KRAS mutation who have pancreatic adenocarcinoma. So patients in this subgroup, so in the high TMB group, only about two-thirds of the patients had KRAS mutation, compared to 92% of the patients with low TMB who had KRAS mutation. So just giving that perspective. So KRAS mutation, which is the most common mutation in pancreatic cancer and is a driver mutation, their rates vary differ from the high TMB group versus the low TMB group.  

And then in addition, in the high TMB group, we found higher rates of BRCA mutation, BRAF mutation, interestingly, and then obviously from the DNA damage repair genes like PALB2 mutation, MSH2 or MSH6, MLH1, and PMS2. So all these mismatch repair protein mutations were higher. As I mentioned before, one-third of the patients with high TMB also had MSI-high. So it's not a totally unexpected finding. I think the biggest finding was that we found more KRAS wild-type pancreatic adenocarcinoma in the high TMB group, almost a third. And those tend to have different targetable mutations like BRCA2, BRAF, and PALB2 mutations. So I think one of the interesting findings is that patients in the high TMB group actually tend to have KRAS wild-type or less KRAS mutations. So they're not necessarily KRAS-driven tumors, and they have a higher chance of having other targetable mutations like BRAF and so on, for which we have therapies for. So it's always something to keep in mind.

Dr. Rafeh Naqash: Would you think that from a DDR perspective, the mutations that you did  identify that were more prevalent in individuals with high TMB, do you think that this is linked to perhaps more DNA damage, more replication stress, more neoantigens leaning toward more tumor mutation burden perhaps? Or is there a different explanation? 

Dr. Amit Mahipal: For sure. As we said, MSI-high tumors have mutations in the DNA damage repair pathway and they definitely tend to have higher TMB. So I don’t think that is very surprising that we found PALB2, or other MMR genes like MSH2, MSH6, MLH1, and PMS2 at much higher rates. I think the interesting finding is the fact that the KRAS wild-type and having BRAF alterations at least that’s not suspected to definitely increase TMB. Although if we look at colorectal cancer, BRAF mutation and MSI are somewhat correlated to patients with BRAF mutations and to have high rates of MSI-high tumors. But that’s not the case in pancreatic cancer. We also found an increase in BRCA2 mutations as well. So I agree that the DNA damage pathway repair gene alteration is not unexpected because they tend to increase TMB, but I think the other mutations were interesting.

Dr. Rafeh Naqash: And I think one other aspect of this, which I’m pretty sure you would’ve thought about is the germline implications for some of these mutations where you could very well end up screening not only the individual patient, but also their family members and have measures in place that we’re trying to enhance screening opportunities there. In your current practice, you are at an academic center but I’m talking about in general with your experience, how common is it to sequence broad sequencing panels in individuals with pancreas cancer? The reason I asked that is I do a lot with lung cancer and even now despite having all those targets in lung cancer which sort of paved the pathway for targeted therapy in many tumor types, we still don’t see a full uptake for NGS Phase I drug development. And I get a lot of referrals from outside and I often see that it’s a limited gene panel. So what is your experience with pancreatic cancer?

Dr. Amit Mahipal: We kind of changed our practice. Similar to you, I’m involved in drug developments. I’ve been a big proponent of NGS for almost a decade now, when didn’t even have targeted therapies but these companies first came in and they’re like, “Okay. We’re very very low chance.” But now obviously, we transformed the treatment for a lot of different cancers. Especially lung cancer, you don’t sometimes even start treatment before you get an NGS panel like you said in situ. So what we’re finding, at least for pancreatic cancer, as you know, the targetable mutations are there but they are somewhat not that common, I would say, in the 10-15% range. So many people would get dissuaded and then it’s like, what’s the point of doing it?  

But I think for those 10% to 15% of the patients, firstly we can really change their treatment course and their prognosis. Secondly, if you don’t do it and they cannot go in a different clinical trials, now we have trials targeting KRAS G12C, but not only that, KRAS G12D which is the most common mutation we see in pancreatic cancer and so on. So it’s becoming very very important. One thing, at least with our practice we adopted last two or three years is sending liquid biopsies or liquid based NGS or blood-based NGS testing. Otherwise, what’s happening I would send a solid tumor NGS from the tissue. And pancreatic cancer as you know has sometimes a very small amount of tissue obtained from FNA. And inevitably after four weeks, we’ll get the result that there’s not enough tumor to do NGS testing. And then the patient comes one or two months later and then we order the test, and that just delays everything. 

So now we adopted a practice where we are trying to send both blood based NGS and solid tumor NGS at the same time the first time of diagnosis when we see the oncologist for the first time. And that has really increased the rate of NGS testing results for our patient population. And it’s not 100%, even in blood-based NGS, sometimes they may not be able to find enough circulating tumor cells to do this blood-based NGS testing, but at least they’re having these. But you’re correct. I think we still see about one third of the patients who had not had NGS testing or referred for phase I clinical trial and have gone through more than two or three line of therapies which is unfortunate for our patients. 

Dr. Rafeh Naqash: That's a very interesting perspective on how important it is to sequence these individuals. As you said, it may not be that all of them may benefit, but the ones that have those important alterations, especially BRCA, PALB, and KRAS could benefit from novel precision medicine-based approaches. 

A question that came to my mind, I saw that you were trying to look at MYC and turmeric low tumors as well. So what is the role of MYC in the context of these individuals? Is there any drug development that’s going on? Because I see small cell lung cancer. MYC is an important target there. These are two different tumors, but it looks like there was a hint of some correlation with respect to some of the findings that you showed. Is that something that you’re currently looking at or planning to look at? 

Dr. Amit Mahipal: I think that if we just talk about MYC in general, it is present at somewhat lower rate. I think we found MYC amplification in about 5% or so of TMB-low patients who had that and not really seen in the TMB-high patients. So right now, I am not aware of any trials targeting MYC in pancreatic cancer. But as you said, if it’s successful in lung cancer, maybe that’s when we can transform into the pancreatic cancer group.

Dr. Rafeh Naqash: Of course we can all learn from each other's specialties.We learned a lot from melanoma with respect to therapy. Hopefully, other fields can also benefit from each other’s experiences in the space of drug development. 

Thank you so much for this interesting discussion. The last few questions are more or less about you as an individual researcher. So could you tell us briefly on your career trajectory and what led you into the space of GI oncology, pancreas cancer, even for that matter, drug development? And some of the advice that you may want to give to listeners who are trainees or early career individuals?

Dr. Amit Mahipal: Sure. So I have gone through some different institutions. During my fellowship, that’s when I really decided that I wanted to do GI oncology. Prior to that, I actually have a Masters in Public Health, where I learned about epidemiological research and how to design clinical trials, how to design cohort studies. My focus was on, actually there was somewhat a lot, but one of my mentors was working on colorectal cancer, and they had this huge database called the Iowa Women's Health Study Database of 100,000 patients. So that's where I started by clearly getting into colorectal cancer and GI cancer in general and how to learn from this database, how to mine these databases, how to do analyses, which seems easy but is actually quite complicated. 

During my fellowship, I think the key to it is finding a good mentor during the fellowship. And I worked with one of the top GI oncologists in the country who’s practicing. And I worked under her and learned a lot not only from the clinic side but also from the research perspective and how sometimes you’ll come up with the ideas during the clinic itself.Like, “Hey, this patient had this and why aren’t we looking into this.” And she would even do some of the therapies based on phase II trials and she was a part of a lot of these trials and learning from those experiences. 

And following my fellowship, I joined Moffitt Cancer Center, where I led the phase I program there. So I was heavily involved in drug development programs, all training programs I’ve been to, NIH in Bethesda, an observership in the CTEP program, and also did the ASCO/AACR Vail workshop, where you really learned a lot in just like one week. So those are kind of opportunities present for fellows and even the early investigators and attendings as well in the first few years can go there, have your proposal. And really they are the world experts in trial design and they’ll talk about how to design trials, how to add collaborators, improve your trial, and basically learn the whole protocol in a week so to speak.  

And then I was at Moffitt Cancer Center for about five, six years. My home was GI so I did both GI oncology as well as phase I. And in terms of the GI oncology, my main focus was pancreatic cancer and liver tumors. Then I was at Mayo Clinic in Rochester for about seven or so years. I kind of did the same thing and solidified my career at GI oncology, looking at liver tumors, and pancreatic cancer and then being a part of the phase I division program. And now, most recently, about a year or so ago, I joined Case Western to lead the GI program here.

Dr. Rafeh Naqash: Are the winters in Cleveland better than the winters in Minnesota?

Dr. Amit Mahipal: For sure. I always say, you don't know cold until you go to Minnesota. It's a different kind of cold. I'm sure people in Dakota might say the same thing, but the cold in Minnesota is very brutal and different compared to any other place I've been to. 

Dr. Rafeh Naqash: Well, it was great learning about you. Thank you so much for spending this time with us and for sharing your work with our journal. We hope you'll continue to do the same in the near future. 

Thank you for listening to JCO Precision Oncology Conversations. Don't forget to give us a rating or review, and be sure to subscribe so you never miss an episode. You can find all ASCO shows at ascopubs.org/podcasts.

Dr. Amit Mahipal: Thank you for having me here, Rafeh. Good luck. Take care.

Dr. Rafeh Naqash: Thank you so much.

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

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

 

Dr. Mahipal:

Consulting or Advisory Role:QED TherapeuticsAstraZeneca/MedImmuneTaiho Oncology

Speakers' Bureau:AstraZeneca

Research Funding:Taiho Pharmaceutical"