JCO Precision Oncology Conversations
Health & Fitness:Medicine
Real-World NTRK Fusions and TRK Inhibitors in Veterans
JCO PO authors Dr. Michael J. Kelley and Dr. Katherine I. Zhou share insights into their JCO PO article, “Real-world Experience With Neurotrophic Tyrosine Receptor Kinase Fusion–positive Tumors and Tropomyosin Receptor Kinase Inhibitors in Veterans.” Host Dr. Rafeh Naqash, Dr. Kelley, and Dr. Zhou discuss the robust Veterans Affairs (VA) National Precision Oncology Program (NPOP), accurate identification of gene fusions, and toxicities landscape of TRK inhibitors. Click here to read the article!
TRANSCRIPT
Dr. Rafeh Naqash: Hello and 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 Stephenson Cancer Center in the Division of Medical Oncology.
Today, I'm thrilled to be joined by Dr. Michael J. Kelley. Dr. Kelley is the executive director of Oncology for the Department of Veterans Affairs. He's also the chief of Hematology-Oncology at the Durham VA Medical Center, and also a Professor of Medicine at the Duke University School of Medicine. And he's also a member of the Duke Cancer Institute. We are also joined by Dr. Katherine I. Zhou who is a Hematology-Oncology fellow at the Duke University. Dr. Zhou also spent time at the Duke Medical Center as part of her fellowship training, which I believe is how this project that was led by her came to fruition.
So thank you both for joining today. This is going to be, hopefully, of very high interest to our listeners and I look forward to chatting with you both.
Dr. Michael Kelley: Great, thanks for having us.
Dr. Katherine Zhou: Thank you for having us.
Dr. Rafeh Naqash: Thank you so much for joining. So I was very intrigued with this paper, and this paper follows a recent podcast that we had with Dr. Alexander Drilon, who's led some of the NTRK tropomyosin receptor kinase inhibitor studies that have been published in the last several years. And we had a very interesting discussion a couple of weeks back and I felt this was going to be a very interesting subsequent discussion into what was also an interesting discussion with Dr. Drilon. So what caught my attention is obviously the fact that you guys in this report, which is a real-world report, did not exactly see what we generally expect from clinical trials as far as response to target therapies in NTRK fusions.
So before I ask you questions related to this project, one of the very interesting things at least I found was the fact is that the Veterans Health Administration is the largest integrated health system. Studies, whether conducted in the UK, for that matter European countries, or in Canada, they have integrated health systems which we do not. But we do have this advantage of the VA trying to do things in a very unique, centralized manner. So I wanted to ask Dr. Kelley first, how is it that you have implemented this National Precision Oncology Program, the NPOP as you call it, into the VA precision medicine workflow and how does it help in conducting research studies like the one that you published in the JCO Precision Oncology?
Dr. Michael Kelley: Yeah, thanks for that question, Dr. Naqash. The NPOP started in 2016 as a national program and right from the beginning it grew out of an effort that was a joint collaboration between both clinical operations in the VA and the Research Office or the Office of Research and Development. It was designed from the very beginning to support discovery, new knowledge generation, and identifying patients for clinical trials in addition to bringing them best-in-class molecular testing and a consultation service.
So it was initially funded out of the Cancer Moonshot 1 in 2016 when President Biden was then Vice President. The VA endorsed the model going forward in 2019 and now it's continued on and grown even bigger, it's expanded both in terms of scope and the complexity of the testing that's been done. So it was offered as services to facilities. They didn't have to do this, but I think they all saw the value of using NPOP to provide this group of services and that's what led to the generation of the robust underlying dataset that Dr. Zhou has used for this paper.
Dr. Rafeh Naqash: Definitely. Thank you so much for that explanation. I did not know, and was not well aware, of how robust this program is. So I think it's a great learning opportunity for our listeners to know that a program like this exists. As we all know, there are different platforms, sequencing platforms, that each institution uses, whether it's commercial or whether it's in-house based. But the fact is, until and unless we have big pool datasets like the ones that you have generated or have access to, it's not easy to answer real-world questions.
So first of all, I'd like to congratulate you and the rest of the VA administration to set up a program like this that hopefully is helping in matching the right patients to the right therapies and in clinical trial approvals. Now, before we take a deeper dive into the study that Dr. Zhou led, I did want to ask you, you have access to this amazing centralized platform, what are the kind of sequencing strategies or platforms that you use as part of this program? And is there an incorporation of molecular tumor boards to help understand some of these sequencing results that sometimes can be a little complicated to understand even for oncologists who look at these reports on a daily basis? So could you tell us a little bit more about that, Dr. Kelley?
Dr. Michael Kelley: Yeah, certainly. So the VA contracts for the sequencing service, currently we're contracting with Foundation Medicine and Tempus for the comprehensive genomic profiling. There are some other services, and before we started using Foundation, there were two other companies that we used. There is a molecular tumor board. Our molecular oncology tumor board is designed primarily for case-based education. But there's also an asynchronous on-demand consultation service that occurs electronically because we have a unified electronic health record system. So any oncology provider in the country can enter a request through what's called an interfacility consult. It comes to a team, that team vets that, discusses it with the appropriate experts; that includes molecular oncologists, molecular pathologists. A lot of oncology pharmacists have been trained at a course that’s at the University of Kentucky.
And we have a lot of experience in doing this since that service was set up in 2016 as well, right from the beginning, because we understood the complexity of the data and the need for every oncologist across our enterprise to have access to the very best interpretation of that.
We also have educational sessions that are integrated into the molecular tumor board time slot we call primers in terms of the underlying science of why you do the interpretations the way you do. And then there's also some additional education that we'll be endeavoring to offer to our staff and our oncologists coming up this year.
Dr. Rafeh Naqash: Excellent. It sounds like you definitely have taken this into a very multidisciplinary approach where you're incorporating oncologists, pharmacists, and perhaps even genetic counselors and then, obviously, keeping the patient at the center and trying to find the best possible therapies that are most relevant for that individual.
Now, going to Dr. Zhou's study here. Dr. Zhou, first of all, it's great to see a fellow lead a study and then especially, I think you're our first fellow on the podcast. We've had a lot of different individuals, but we have not had a fellow before. So thanks for coming.
Could you tell us, for our listeners, what drove your interest into NTRK fusions? As we know, they are rare, something that is not commonly seen, and we do have clinical trial data in this space. So what was the idea behind looking at a real-world data set? Did you start out with a hypothesis or were you just interested to see how targeting these fusions in the real-world setting, actually, what kind of results does it lead to?
Dr. Katherine Zhou: Yeah, well, first of all, thanks for the question. And I do just want to mention that although I did sort of bring this project to the finish line, it was started by another fellow, Vishal Vashistha. So just wanted to mention that. And I think the interest was really just that NTRK is such a rare fusion and just a difficult one to be able to study, like you said, in the real-world setting. And we have the advantage of having so much data through the VA and through NPOP, specifically. And so having seen such great results with the TRK inhibitors and clinical trials, I think there's this big question of how that translates into the real-world setting. We have the ability to do that with our large patient population.
Dr. Rafeh Naqash: Excellent. And again, it's nice to acknowledge the support that you had from the other individual who co-led this study. Now, since you would have, I’m guessing, done most of the analysis here and looked into the whole idea of the kind of results that you saw—and from my understanding, you looked at the entire VA data set and tried to understand first the incidence or frequency of NTRK fusions and also responses to treatment, which I think is the main message—but could you tell us a little bit more about the data set? How did you acquire the data set, and what it took to analyze? Because obviously every project has a very unique story, and I'm guessing there's one very unique story here, since as a fellow you have limited time to do all this interesting work. So how did you navigate that and analyze and work with some of the things that you had to look at to get to the results?
Dr. Katherine Zhou: Yeah, so again, this was work that was done with multiple people involved, of course. And we used what we had, the resources we had available, some tools we had available through the VA. So first, looking at NPOP and looking at patients who are sequenced through NPOP, we could just find all the ones who had an NTRK rearrangement of some kind. The second way we went about finding patients was through the CDW or the Corporate Data Warehouse where we could see which patients were prescribed larotrectinib or entrectinib and kind of go backwards from there and see which of those patients had NTRK alterations or specifically NTRK rearrangements. And so we combined the patients from both of those different methods to come up with our cohort at the end of 33 patients with NTRK rearrangements and 12 patients who are treated with TRK inhibitors.
Dr. Rafeh Naqash: Excellent. Could you walk us through what was the subsequent analysis as far as how many NTRK fusions? I know you mentioned in the paper about DNA versus RNA-based testing. So how many were DNA-based, how many were RNA-based? I think there's some element of ctDNA-based testing also, or what tumor types those people had so that we get an understanding of what's the landscape of the findings that you had.
Dr. Katherine Zhou: Sure. Since this is a real-world setting, as you may expect, the vast majority of the sequencing was done through tissue DNA sequencing, and that was the case. So for the 25 patients who were sequenced through NPOP that we found who had NTRK rearrangements, 23 of them had tissue DNA sequencing. And then one was tissue DNA RNA, and one was cell-free DNA sequencing. And so using that and being able to go back and look at how many patients have been sequenced in NPOP in total, we could kind of come up with a yield, although the numbers are very small. But we do see that there does seem to be probably a lower yield, for example, with cell-free DNA sequencing, as one might expect.
And then looking at our total group of 33 patients, if we look at what types of cancers they had, we did have quite a few patients just based on prevalent tumors at the VA, I think, and in the population, prostate cancer was common, lung cancer, and then we had smaller numbers of colon and bladder, and I think there's a pancreatic cancer patient. We did have some of these rarer tumor types that more commonly have NTRK fusions as well, so like papillary thyroid carcinoma, and salivary gland cancers as well as soft tissue sarcomas.
Dr. Rafeh Naqash: Question for you, Dr. Kelley, related to this data set: do you think that given that the denominator that you have is a unique population, the VA population, that's often males, they're usually above the age of 18, could the frequency have been influenced by that denominator where you may not have been able to capture, let's say, some of the rarer tumors that happen in the younger patient population, for that matter? Could that be a little bit of a bias here?
Dr. Michael Kelley: Definitely. The population of veterans that have cancer that is treated in the Veterans Health Administration do represent generally adult males in the United States, but there is some skewing in certain regards. One of them is towards a higher frequency of smoking status. So not current smoking, which is actually about the same as the national average of about 11%, but the former smoking rate is about twice as high as it is in the rest of the United States. So we may have a lower frequency of some actionable variants in cancers in general because there's a higher etiological role for tobacco smoke in our population. But overall, looking at adult men if we look at like EGFR mutations, our incidence of EGFR mutations in adenocarcinoma is similar to what is reported in other real-world evidence bases from the United States, which is significantly lower than that which is found in academic medical centers.
Dr. Rafeh Naqash: Thank you. I'm a big fan personally of real-world data sets. I do a lot of this with some other collaborators and generally, I do phase I trials, which is why I’m interested in precision medicine. And two weeks back, actually, I had a patient with prostate cancer, who ended up having NTRK fusion on a liquid biopsy. Now, you do talk about some of this related to in-frame or out-of-frame fusions and how that can have interesting aspects related to the kinase domain functionality and RNA expression. Dr. Zhou, for the sake of our listeners, could you briefly describe why understanding some of that is important and what implications it has?
Dr. Katherine Zhou: Yeah, so I think the oncogenic NTRK fusion that we think of and that's being targeted by the TRK inhibitors is a fusion 5-prime of a protein that forms a dimer and on the 3-prime end is the kinase domain of the tropomyosin receptor kinase. And so you have to have some kind of a gene fusion that results in not only the transcription of that RNA fusion, RNA transcript, but then the translation of that fusion protein. So that needs to be, like you mentioned, that has to be in frame so that the entire protein is translated and expressed and it needs to include the kinase domain. It can't be the other end of the NTRK gene. And both of the genes need to be in the same orientation, of course. And then also the partner gene probably matters in that the ones that we know that actually cause activation of this oncogene are the ones that sort of spontaneously dimerize. And so that's a lot of requirements that we don't necessarily see when we just get, for example, a DNA sequencing result that says there's an NTRK rearrangement.
Dr. Rafeh Naqash: Excellent way to describe the importance of understanding the functionality of the activated oncogenic fusion. Now, I know here in most of the patients that you have is DNA sequencing and I'm sure you'll talk about some of the results. And when you connect the results to the kind of data that you have, do you think not having the RNA assessment played a role in not knowing perhaps whether those fusions were functionally active?
Dr. Katherine Zhou: Yes, I think we can't know for sure without having the RNA sequencing data. But certainly, that is a pattern in our small number of patients that we saw and something that makes sense just in terms of the mechanism of this oncogenic fusion protein. So I think that is a question of when should we be doing RNA sequencing to confirm that a fusion that we see on DNA sequencing is actually transcribed into RNA and how do we use RNA sequencing in a cost-effective and useful way to be able to detect more of these NTRK fusions that are actually clinically relevant.
Dr. Rafeh Naqash: I absolutely agree with you and this is an ongoing debate. I know some platforms, commercial platforms that is, have incorporated RNA sequencing both bulk or whole transcriptome as part of their platform assessments, but it's still not made inroads into some other sequencing platforms that are commercially used. So it's an ongoing debate, but at the same time helping people understand that certain fusions need some level of RNA assessments to understand whether they're functionally active or not. Which again has implications, as you pointed out in terms of therapies are extremely relevant.
Now, going to the results, which again was very interesting, could you tell us about the findings from the therapeutic standpoint that you observed and what your thoughts are about why you saw those results which were very different from what one would have expected?
Dr. Katherine Zhou: Right. So in the clinical trials of larotrectinib and entrectinib, there were quite high objective response rates on the order of 60%, 70%, even almost 80%. In our very small real-world group of 12 patients who were treated with TRK inhibitors, nobody had an objective response and five patients had stable disease and everybody else, the other seven patients, progressed.
And so the question is why did we see such a big difference compared to the trials? I sort of think of this as two big buckets. One is the population that we were looking at. So this is a real-world population. For example, in the clinical trials, there were almost no Black or African American patients, whereas here we had about 30%-40% Black or African American patients. Because it’s a VA population, it was very heavily male, of course, the age groups are also different in that we didn't have children in the VA population whereas children were included in the trials. And the tumor types also differed because I think in the trials, which makes sense, there's a bias towards tumor types that have more NTRK fusions, and some of the tumor types we were looking at are just common tumor types like prostate and lung cancer where NTRK fusions are not common. But just because there are so many patients with these cancers, we did see them. And so certain of these groups, particularly certain racial and ethnic groups as well as certain tumor types, were not really represented in the trial to the extent that we can make conclusions about whether TRK inhibitors are effective in this population. So that's one.
The second part, I think we've already talked about some, is just the method of detecting these NTRK fusions and how many of these NTRK fusions were actually truly producing oncogenic fusion proteins. And I tried to sort of categorize some of these fusions as being canonical in that they've been more studied. We know the partner gene, they are known to produce an oncogenic protein and to respond to TRK inhibitors. But actually of the four patients who had what we called canonical fusions, all four of them had stable disease at least, whereas the ones that were noncanonical mostly did not have a response or have even stable disease and mostly just progressed. And so then you wonder whether they even had the actual target protein we thought we were targeting. So this is where the real-world setting we're not doing the RNA sequencing or this additional testing to confirm that it's an oncogenic fusion protein.
Dr. Rafeh Naqash: And I do see in your results there's a patient especially—you pointed out canonical and noncanonical fusions—you have a patient with a papillary thyroid cancer that I believe had a stable disease for close to two years plus. Is there anything interesting apart from an NTRK fusion in that specific patient where certain co-mutation could have played a role or certain other factors that do you think played into the fact that this patient had stable disease but didn't respond on the TRK inhibitor?
Dr. Katherine Zhou: I don't have a great answer for that. I think this is one of the cancers that was well represented in the trials and that commonly has NTRK, or more likely has NTRK fusions. And this was a well-studied canonical NTRK fusion. So I think those are all reasons. The question of co-mutations I think is really interesting. We didn't have the data for every single patient, but for the ones we looked at a lot of the time, NTRK fusions are mutually exclusive with other driver mutations. So we didn't see a whole lot of commutations that we could sort of differentiate between responders or stable disease and progression.
Dr. Rafeh Naqash: Thank you. Going to the toxicities, as a phase I trialist myself toxicity is the bane of my existence where we have to label toxicities, attribute toxicities, understand toxicities. The trial, obviously, as you very well know, that in the trials, they didn't have a lot of toxicities that caused patients to come off or required significant dose reductions, which is not the case compared to what you saw. Could you tell us a little bit about the landscape of toxicities for TRK inhibitors and what you saw in your cohort? That, again, I feel was interesting.
Dr. Katherine Zhou: Of the 12 patients, I think two-thirds of them had either dose reduction or interruption or discontinuation, or some combination of the above. The toxicities we saw were more common than, or at least led to discontinuation and interruption and dose reduction more commonly than in the trials. But the toxicities we saw were also seen in the clinical trials. So LFT elevations, creatinine elevations, neurotoxicity, some cytopenias. We didn't actually see a whole lot of that, but those were present as well, and then some sort of nonspecific things like fatigue. And so, as much as we could tell from retrospective trial review, at least these were severe enough to lead to holding the drug.
Dr. Rafeh Naqash: Thank you so much, Dr. Zhou. Question for you, Dr. Kelley. Putting this into perspective, the analysis that you did, how would you connect it to other real-world questions that one could answer using these kinds of data sets? So basically, what are the lessons learned from this amazing program that you guys have run successfully and are, I'm guessing, expanding in different directions? And how can you use a program like this to look at some of these unique questions using real-world data sets?
Dr. Michael Kelley: There are a couple of, I guess, next steps for us that are based off this study and other information that we've gotten in other analyses from our NPOP data set. So, first of all, access to an RNA-Seq test. So that has been resolved to some extent, in that we now have two options for comprehensive genomic profiling, one of which does have RNA-Seq.
And then the other approach that we're doing is to do more robust data generation. So we're going to be launching a study to collect prospective data on patients who are treated with off-label drugs. And as part of that, we will also have an on-label cohort for rare populations or any investigator in the VA who's interested in a particular drug or a particular genetic variant. They'll be able to tie into this protocol, and we will then collect data from across the system prospectively, which we think will improve the quality to some degree.
And then thirdly, I think there's an opportunity to merge the initial generation of data in rare genetic types or other populations, which are highly selected by doing a distributed type of clinical trial where patients can be enrolled in prospective treatment trials. So we're not just generating data based on their real-world exposure to FDA-approved drugs, but we're generating data as we're developing the new drugs, we can have a much more heterogeneous and representative population of patients enrolled in clinical trials. So this is called the decentralized clinical trial model. We're starting to launch some trials with industry partners in this area to test out the model. If it works, I think we'll be able to help contribute to the knowledge that we all can use in terms of the patient types, the patient characteristics, but also some of the different tumor characteristics, and also to bring clinical trial opportunities to a more representative group. A lot of the initial clinical trials are done in urban areas, rural populations in VA are about a third of our patients live in rural areas, compared to only 14% of the country. So we think this is a very important diversity issue that should be addressed. Those are some of the ways that we're taking a lesson from this trial and other data that we have to sort of bring it forward.
Dr. Rafeh Naqash: Those are excellent next steps and I think the kind of work that the VA is doing and this specific program, Precision Oncology Program, the NPOP program is doing, it's definitely setting up a unique standard in the United States where we have been limited by not having a centralized database. So setting something up of this sort hopefully will help answer a lot of these unique, interesting questions as you have access to data. And then the fact that you mentioned decentralized clinical trials and trying to cater to this access issue for patients in the VA system, I think that would be huge.
And again, I congratulate you and your team on these efforts, and once again, thank you for joining us today and making JCO Precision Oncology a destination for your interesting work. We hope to see more of this work subsequently and hopefully, I get a chance to talk to you more about all the exciting stuff that you guys are leading within the VA health system.
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 asco.org/podcast.
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.
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.
Bios:
Michael J. Kelley, MD, is Executive Director of Oncology for the Department of Veterans Affairs, Chief of Hematology-Oncology, Durham VA Medical Center, Professor of Medicine at Duke University School of Medicine and Member of the Duke Cancer Institute.
Katherine I. Zhou, MD, PhD is a hematology-oncology fellow at Duke University. She also spends time at the Durham VA Medical Center as part of her fellowship training.
COIs:
Michael J. Kelley, MD
Research Funding: Novartis (Inst), Bristol-Myers Squibb (Inst), Regeneron (Inst), Genentech (Inst), EQRx (Inst)
Katherine I. Zhou, MD, PhD: No disclosures
Create your
podcast in
minutes
It is Free