JCO PO author Dr. Brandon Huffman shares insights into his JCO PO article, “Analysis of Circulating Tumor DNA to Predict Risk of Recurrence in Patients With Esophageal and Gastric Cancers” and discusses the article’s findings of ctDNA levels in the preoperative, postoperative, and surveillance settings in patients with EGC. Host Dr. Rafeh Naqash and Dr. Huffman discuss ctDNA assessments, treatment paradigms and interventions, and tumor-informed assays.

TRANSCRIPT

Dr. Abdul Rafeh Naqash: Hello, and welcome to J CO 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 I'm also an Assistant Professor in Medical Oncology at the OU Stephenson Cancer Center. Today, I am excited to be joined by Dr. Brandon Huffman. Dr. Huffman is a gastrointestinal medical oncologist, and he's also an instructor in medicine at the Dana-Farber Cancer Institute at the Harvard Medical School. He's the lead author on today's JCO Precision article, "Analysis of Circulating Tumor DNA to Predict Risk of Recurrence in Patients with Esophageal and Gastric Cancers."

Our guest’s disclosures will be linked in the transcript.

Dr. Huffman, welcome to our podcast and thanks for joining us today.

Dr. Brandon Huffman: Of course. Thanks for having me.

Dr. Abdul Rafeh Naqash: For the sake of this discussion, we'll refer to each other using our first names. So, Brandon, exciting to have you today. We're going to talk about this very interesting topic on circulating tumor DNA and how your team used the ctDNA assessment in patients with esophageal and gastric cancers. For the sake of the listeners, could we start by asking you what are the current treatment paradigms for early-stage esophagogastric cancers? Since you practice this on a daily basis, what is the current approach, briefly, which will play into how this study looked at ctDNA in the context of early-stage esophagogastric cancers?

Dr. Brandon Huffman: Yes, definitely. Thanks first for having me. Thanks for highlighting our work, and I'm really excited to talk with you about our manuscript and research today. To answer your question about how to treat localized esophagogastric cancer, it's a little bit more specific depending on where in the esophagus, GE junction or stomach where the tumors arise. For instance, we treat esophageal and upper gastroesophageal junction cancers with, often, chemoradiation, neoadjuvantly, and that is followed by surgery. And if there's a pathologic incomplete response, then many patients will get Adjuvant Nivolumab, a PD-one inhibitor, whereas the lower the tumor is in the upper GI tract, most often, perioperative chemotherapy is used for the lower GEJ and gastric cancers.

Dr. Abdul Rafeh Naqash: Thank you so much. And I know, I think to some extent, if I remember correctly, immunotherapy has been incorporated into this paradigm. Is that a fair assessment?

Dr. Brandon Huffman: That's exactly right. So, excitingly, we treat patients with neoadjuvant chemo or chemoradiation, and surgery is really the crux of the treatment paradigm for esophagogastric cancers in general. However, recently the CheckMate 577 clinical trial for the use of adjuvant Navolumab showed an improvement in disease-free survival in patients who had an incomplete path response. They used one year of Nivolumab compared to placebo. So it has recently become a standard of care where I practice, and I feel like a common practice around the country.

Dr. Abdul Rafeh Naqash: Thank you so much. Now, going to the premise of this paper where you and your team basically looked at circulating tumor DNA as a prognostic marker in these patients that had early-stage esophagogastric cancers, was there a specific reason why you wanted to look at the early stage? What was the rationale for evaluating this biomarker in this patient population?

Dr. Brandon Huffman: So, esophageal and gastric cancers affect a large number of patients every year. And unfortunately, despite our best efforts with curative intent therapy, over 50% recur within three years. So we know that there are pre surgical risk factors such as a larger bulky primary tumor or lymph node-positive disease that increase the risk for progression or recurrence after surgery. And we know, in addition, in other GI malignancies and other malignancies such as colorectal cancer, for instance, that the presence of circulating tumor DNA after surgical resection of localized tumors is associated with an increased risk of recurrence. So this has actually led to clinical trials investigating whether or not ctDNA can be integrated into the decision-making for adjuvant colorectal cancer treatment, such as ongoing trials such as the BESPOKE trial, COBRA, DYNAMIC trials that have recently been reported.

The use of ctDNA is being used in other malignancies. And to give you a little bit of background, this project started when I was seeing patients with Dr. Sam Klempner at Mass General during my fellowship, where I was in the combined Dana-Farber/Mass General program. And he and others had begun collecting serial plasma samples on every patient we saw with esophagus, gastroesophageal junction and gastric cancers to assess for the presence or absence of ctDNA. And we used the tumor-informed ctDNA assay from Signatera, which, for those who aren't familiar, this is a ctDNA platform where a panel is built from the results of whole exome sequencing on the patient's FFPE tumor. The panel includes 16 patient-specific somatic single nucleotide variants for each patient, and it's new for each patient. Once that panel is built, the cell free DNA is tested from a plasma sample. And if there are two or more of the tumor-specific variants present, then they're considered ctDNA positive. So some of those colorectal cancer trials that I mentioned before are using this assay, and we wanted to investigate whether or not this high-risk population could be further assessed for risk of recurrence.

Dr. Abdul Rafeh Naqash: Excellent. Thank you so much. And I know that a lot of these ctDNA based assessments have made inroads into the GI malignancy space, lesser in the other tumor types. I think we are all trying to catch up to what you guys are doing in the early-stage colon cancer space or the early-stage esophagogastric cancer space. So it's definitely very a interesting avenue to assess minimal residual or molecular residual disease.

Now, going back to the methodology, I found it very interesting, and I think it's very important for listeners especially to understand the context of ctDNA assessments because I think a majority of oncologists are used to the liquid biopsy aspect. But this is not necessarily the liquid biopsy. It's somewhat different. So what I've understood, and I'd like to ask you to explain in the context of tumor-informed and tumor uninformed assays, what are the assays that are available, and how do they differ in terms of serial monitoring? And why is this ctDNA-based assessment somewhat different or more patient-customizable than our regular liquid biopsy assays, which are also blood based but not tumor-informed?

Dr. Brandon Huffman: That is the question of the hour. And many different research projects are ongoing to try and identify which one is better, if one is better. I know that there are some commercial assays, for instance, that are not tumor-informed. They take a blood sample and then test for cell free DNA. The risk behind that is it's testing for common genetic mutations from a next-generation sequencing panel platform. And it may also detect CHIP variants or clonal hematopoiesis of indeterminate potential variants that aren't related to the underlying solid tumor malignancy.

So a tumor-informed assay, for instance, such as the one that we used in this study, uses the patient's tumor and sequences it with whole exome sequencing and identifies very specific variants within the tumor that are only present within the tumor because they compare it also with a normal blood sample from the patient at the same time. And so they pick tumor-informed specific variants that then they test for on their assay. And that increases the sensitivity of the ctDNA assay so that you can really try to understand, is this cell free DNA that we are detecting related to the tumor or can we ignore it potentially? I don't know if we can necessarily ignore it in all honesty because it could affect- there's a lot of ongoing work that is looking at the risk of CHIP. But overall, this is specific for the primary tumor that we were investigating.

Dr. Abdul Rafeh Naqash: I definitely agree with you there. I think the important point, as you mentioned, is that using the whole exome approach, in the blood and the tumor, you're able to eliminate the CHIP variants or the germline variants that may not be contributing. And that way you're able to specifically look at certain genetic alterations that eventually, I think using PCR-based approaches, you identify the same and quantify the same in the blood serially. And that's how this tumor-informed assay is somewhat unique and different.

Now, going to the crux of this study, could you tell us a little about the patient population? I think you stratified patients. You had a pre-operative cohort, you had an MRD cohort, you had a surveillance cohort, and you had a cohort where you assessed ctDNA positivity at any time point. So, several different cohorts, and you assess recurrence-free survival in those cohorts. Could you tell us a little bit more about how you evaluated these cohorts? What were the selection criteria, and how many patient samples did you have for these different cohorts?

Dr. Brandon Huffman: Absolutely. So, we aimed to determine the feasibility of testing ctDNA in patients with gastroesophageal cancer. And so, there were several clinicians from over 70 institutions across the United States who began prospectively collecting serial plasma samples for the presence or absence of the tumor-informed ctDNA. And they included patients from stages one through stage four, gastroesophageal cancer specifically, they included patients who were stages one through four with gastroesophageal cancer. They were collected at the discretion of the ordering clinicians and then incorporated into their routine clinical care as they saw fit. Within this dataset, we have a subset, a large number of patients that is unique to this dataset, specifically in that we have clinical outcomes, treatment, and follow-up data for the patients that were reported on the main findings in the paper. So, overall, we collected and analyzed over 900 plasma samples in almost 300 patients with gastroesophageal cancer, esophageal, gastroesophageal junction and gastric cancers. And in many of the analyses, we lumped them all together. But then we also wanted to separate it out because, as I mentioned before, the treatment paradigm does differ amongst a more proximal esophageal tumor compared to a distal gastric cancer.

So, we focused a majority of our analyses on the detection of ctDNA and localized disease, which included 212 patients with stages one through three gastroesophageal cancer. And I would say we had three major findings. Most of the patients who were tested beforehand, which was a small subset, as I mentioned, this was pragmatic at the discretion of the ordering clinician, but most of the patients who were tested beforehand had positive preoperative ctDNA present. Of the patients who were tested for postoperative ctDNA at any time point, and then specifically within the different subsets of populations that we talked about, postoperative ctDNA was associated with at least a tenfold increased risk of recurrence in all subsets. And ctDNA detection postoperatively was independently associated with recurrence when controlling for age, sex, tumor location, and microsatellite status.

So, a few of the populations that we wanted to test for, one in particular was the molecular residual disease, or MRD window. We labeled this MRD window as the time from surgical resection until 16 weeks. So, if patients were ctDNA positive within that window, we counted that in the primary outcome. And the reason that we chose the MRD window, in addition to this time point of 16 weeks - I should say that the 16 weeks is without any therapy postoperatively, so they have not been treated with any chemo or immunotherapy in this window. We thought that this MRD window was an interesting research topic because the CheckMate 577 Adjuvant Nivolumab clinical trial identified that 16 weeks was the window in which patients could be enrolled up until that timepoint to receive adjuvant nivolumab. So, we're thinking from a future project standpoint, a future clinical trial, perhaps, that if we have identified that patients who are ctDNA positive within this timepoint window, is there an increased risk for recurrence? Because if there is, then perhaps nivolumab intervention will decrease that risk or something that is escalated further. And that's a question that we don't have the answer for, a question that our data can't answer adequately. But it's an interesting one that I see the future questions that can be answered from these data.

Dr. Abdul Rafeh Naqash: Thank you so much. And I agree with you there that this is a very intriguing approach of finding out whether treatment escalation has to be done based on ctDNA positivity, but also, conversely, treatment de-escalation, which there is a lot of emphasis being laid on, especially in the early phase trial in lung cancer, especially in the early setting when targeted therapies or immunotherapies are approved for one to three years, depending on what kind of therapy you're looking at. In those individuals that perhaps have negative ctDNA after one year, maybe therapy de-escalation would be a reasonable approach. So, definitely more interesting clinical trial ideas in this space focusing on ctDNA assessments.

Now, one of the questions that comes to my mind is, when you use ctDNA-based assessments, initially, the patient gets biopsied, and it usually takes four to six weeks for ctDNA-based assessments to come back--I'm talking about tumor-informed assay results to come back, in my personal experience. So, could that potentially, or in your practice, how do you mitigate those delays? If you're trying to schedule a patient for surgery, for example, does that cause any delays in any care because you're trying to get the assessment done, or does your workflow proceed as planned and then you get the results and then subsequently you perhaps make a decision based on their ctDNA assessment?

Dr. Brandon Huffman: At the present time, we are trying to gather more data to understand what we should do with the results that we're receiving. And I think the starting point of collecting serially to just understand the process is helpful. One of the questions we wanted to know that we weren't able to answer with this dataset was: is there lead time? In many cases, ctDNA detection can occur even a year prior to radiographic recurrence. In our case, because this was a pragmatic, clinically at the discretion of the investigators when they decided to test patients for ctDNA, there is heterogeneity among those who are ctDNA positive, and when they get their radiographic imaging, maybe they were moved up. I know in our practice with Dr. Klempner, when I was seeing these patients with him, it was a flag for us to order scans earlier in a patient that we might not have historically ordered so that we could then see, is there something intervenable? Maybe there was a positive lymph node on PET imaging that we could radiate or that wasn't included in the neoadjuvant radiation, for instance. So, we could not predict the lead time from positivity to radiographic recurrence, but I think that that's the hope is that we detected micrometastatic disease, my hope is that we can intervene in the future. But these data aren't able to quite answer that question perfectly.

Dr. Abdul Rafeh Naqash: Sure. And there's definitely caveats to doing this in a pragmatic manner based on investigator assessments.

Now, another question I was thinking of is, when you do do these ctDNA based assessments, and since these are tumor-informed, meaning you biopsy the tumor initially, you identify certain single nucleotide variants and those are the ones that you basically barcode and do PCR assessments using blood. We've learned time and again that tumors can change based on the kind of therapy that you give the patient. So, if your tumor is seeing FOLFOX nivolumab, or all the other novel therapies that you guys give in the setting, is there a chance that the tumor changes over time and you may not be able to capture those newer single-nucleotide variants that are coming up? It's just a provocative question, but I wanted to see what your thoughts are on that.

Dr. Brandon Huffman: It's a great question. I don't entirely know the answer. I'll just be forthright about that. I do think that when designing these assays, they try to choose the more clonal rather than subclonal variants. And so the hope is that, despite the heterogeneity that we know occurs in esophagogastric cancers, we can eliminate that possibility. But you're right, there's no perfect way of knowing that.

Dr. Abdul Rafeh Naqash: I really appreciate you using that word subclonal versus clonal. I think that perhaps makes a difference there. But again, more to do in this field to understand how the tumor evolves and whether it's the clonal mutation, subclonal mutation that needs to be followed. But definitely a lot of interesting work in this space that's ongoing, and, like you mentioned, there are ongoing trials, and both in the neoadjuvant adjuvant space, this field is definitely moving fast in the right direction.

I briefly want you to highlight that one patient case study example that you had. And this was a patient with oligometastatic disease recurrence where you used the ctDNA assessment. And I do some of this in my daily practice, and I really found it useful to have this sort of a patient case example that elaborates in the bigger picture of how this kind of assessment works in a real-life scenario. So it's not just data, it's a patient's trajectory over the course of time where the treating physician was able to use this assay. Could you tell us a little bit more about this individual example here?

Dr. Brandon Huffman: Yeah, absolutely. So this was a 56-year-old man that we saw in clinic with stage three esophageal adenocarcinoma, and was treated with the standard neoadjuvant cross chemoradiation, had an R0 resection with residual disease with a significant treatment effect. And there were lymph nodes that were positive on surgical resection with 39 lymph nodes removed. The patient recovered well and was followed with the standard of care radiographic and clinical surveillance. We also were looking for ctDNA, and what we noted was that there were, often you find these undulating pulmonary nodules that come and go, and they may or may not be infectious, and maybe there's one that's sub-centimeter that slowly grows, and what we found was that at about five months post-surgery, there was a positive ctDNA MRD, and we repeated it at short interval and noted a rising value, which this assay will give you a quantitative value. Once we did that, we ordered imaging and saw a nine-millimeter pulmonary nodule and ultimately biopsied it. It was there in the right upper lobe, and it was positive for metastatic adenocarcinoma. So we treated the patient with the standard FOLFOX plus Nivolumab and actually did SBRT, stereotactic body radiotherapy, to the lung metastasis. And his ctDNA became undetectable.

So because FOLFOX is toxic, we transitioned to a maintenance of Nivolumab and he was on maintenance therapy for several months and had no radiographic evidence of disease and remained ctDNA negative for twelve months. So we biopsied the right upper lobe lung lesion. It was positive for metastatic adenocarcinoma, and then after a multidisciplinary discussion, we treated him with SBRT and then FOLFOX and Nivolumab and then dropped down to Nivolumab maintenance once his ctDNA was undetectable. That highlights the fact that this was an isolated recurrence, which we continued to monitor, and then he had another site of disease a few months later, and we did SBRT to that area while he maintained on just Nivolumab, and the ctDNA came down as well.

So I think, although it doesn't prove anything necessarily, other than demonstrating there is a correlation with the newly diagnosed metastatic disease, it does note that you can use this in dynamic ways, and if it really helps patients live longer, although this is anecdotal--who knows? If we hadn't done SBRT to that area, it was 9 mm. We could have waited until it grew, but then maybe some subclonal, more aggressive metastasis could have really put this patient in a much tougher situation. So it's an interesting case example, and there are several others that we could have put in here that are pretty similar.

Dr. Abdul Rafeh Naqash: Thank you so much for highlighting that case. I couldn't agree more that there is a certain aspect to the ctDNA assessment, where in individuals like the example that you've highlighted here, this can provide lead time potentially and help with earlier management of perhaps more like oligometastatic disease rather than diffuse disease burden.

And in that context, one of the questions that I was going to ask you, based on your data, was there any correlation of tumor burden preoperatively and ctDNA positivity after surgery that you guys were able to identify or thinking of identifying?

Dr. Brandon Huffman: Unfortunately, with our data set, we weren't able to look at that assessment of comparing the overall tumor burden to the quantitative value. But it's an interesting one because we know that in other malignancies, for instance, if there is a correlation of overall disease burden, it also depends on the tumor type, but we also know that perhaps patients will respond differently to chemo or immunotherapy if they have a lower tumor burden, if they have a lower ctDNA value, potentially. I think that's an interesting question for a future project.

Dr. Abdul Rafeh Naqash: Thank you so much, Brandon.

We do like to talk a little bit about the person behind the work. So tell us a little bit more about yourself, your training, your interests, and some little advice for other early-career investigators who might be looking into a similar space and hopefully get inspired by the kind of work that you've done or are planning to do.

Dr. Brandon Huffman: Sure. So, as I mentioned, when I started this project, I was in fellowship. I was seeing patients with Dr. Sam Klempner at Mass General, where I saw patients with him for a year, and as part of my clinical training in the Dana-Farber/Mass General HemOnc Fellowship. Since that time, I have graduated fellowship. I'm a GI Medical Oncologist at Dana-Farber Cancer Institute, and in the GI division, I see patients with all GI malignancies, and I focus on the development of clinical trials in upper GI malignancies, along with investigating the use of circulating tumor DNA as a biomarker, hopefully, we can understand whether it's a predictive biomarker that we can intervene upon in the future.

I think the greatest advice that I received and that I will give to all future trainees; I'm not sure that I'm qualified to tell this to all the junior investigators, but here it is: Find yourself a mentor who really cares and invests in you and your ideas. I have that with Sam, and this project was an incredible part of my development as a junior investigator. I've asked really interesting questions. There are more questions that can be answered from this data set, and I'm excited for the opportunity.

Dr. Abdul Rafeh Naqash: Thank you so much, Brandon. Thanks for taking the time to speak with us today and thank you for choosing JCO Precision Oncology as a destination for your work. Hopefully, we'll see more of this subsequently in the years to come.

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

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.

Guest Bio

Dr. Brandon Huffman, MD, is a gastrointestinal medical oncologist and Instructor in Medicine at Dana-Farber Cancer Institute and Harvard Medical School.

Guest Disclosures

Brandon M. Huffman

Stock and Other Ownership Interests: Doximity

ctDNA for Recurrence Prediction in Esophagogastric Cancers

JCO Precision Oncology Conversations