ctDNA in Metastatic Invasive Lobular Carcinoma
JCO PO author Dr. Foldi at UPMC Hillman Cancer Center and University of Pittsburgh School of Medicine shares insights into the JCO PO article, "Personalized Circulating Tumor DNA Testing for Detection of Progression and Treatment Response Monitoring in Patients With Metastatic Invasive Lobular Carcinoma of the Breast." Host Dr. Rafeh Naqash and Dr. Foldi discuss how serial ctDNA testing in patients with mILC is feasible and may enable personalized surveillance and real-time therapeutic monitoring. 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 am your host, Dr. Rafeh Naqash, podcast editor for JCO Precision Oncology and Associate Professor at the OU Health Stephenson Cancer Center at the University of Oklahoma. Today, we are thrilled to be joined by Dr. Julia Foldi, Assistant Professor of Medicine in the Division of Hematology-Oncology at University of Pittsburgh School of Medicine and the Magee-Womens Hospital of the UPMC. She is also the lead and corresponding author of the JCO Precision Oncology article entitled "Personalized Circulating Tumor DNA Testing for Detection of Progression and Treatment Response Monitoring in Patients with Metastatic Invasive Lobular Carcinoma of the Breast." At the time of this recording, our guest's disclosures will be linked in the transcript. Julia, welcome to our podcast, and thank you for joining us today. Dr. Julia Foldi: Thank you so much for having me. It is a pleasure. Dr. Rafeh Naqash: Again, your manuscript and project address a few interesting things, so we will start with the basics, since we have a broad audience that comprises trainees, community oncologists, and obviously precision medicine experts as well. So, let us start with invasive lobular breast carcinoma. I have been out of fellowship for several years now, and I do not know much about invasive lobular carcinoma. Could you tell us what it is, what some of the genomic characteristics are, why it is different, and why it is important to have a different way to understand disease biology and track disease status with this type of breast cancer? Dr. Julia Foldi: Yes, thank you for that question. It is really important to frame this study. So, lobular breast cancers, which we shorten to ILC, are the second most common histologic subtype of breast cancer after ductal breast cancers. ILC makes up about 10 to 15 percent of all breast cancers, so it is relatively rare, but in the big scheme of things, because breast cancer is so common, this represents actually over 40,000 new diagnoses a year in the US of lobular breast cancers. What is unique about ILC is it is characterized by loss of an adhesion molecule, E-cadherin. It is encoded by the CDH1 gene. What it does is these tumors tend to form discohesive, single-file patterns and infiltrate into the tumor stroma, as opposed to ductal cancers, which generally form more cohesive masses. As we generally explain to patients, ductal cancers tend to form lumps, while lobular cancers often are not palpable because they infiltrate into the stroma. This creates several challenges, particularly when it comes to imaging. In the diagnostic setting, we know that mammograms and ultrasounds have less sensitivity to detect lobular versus ductal breast cancer. When it comes to the metastatic setting, conventional imaging techniques like CT scans have less sensitivity to detect lobular lesions often. One other unique characteristic of ILC is that these tumors tend to have lower proliferation rates. Because our glucose-based PET scans depend on glucose uptake of proliferating cells, often these tumors also are not avid on conventional FDG-PET scans. It is a challenge for us to monitor these patients as they go through treatment. If you think about the metastatic setting, we start a new treatment, we image people every three to four cycles, about every three months, and we combine the imaging results with clinical assessment and tumor markers to decide if the treatment is working. But if your imaging is not reliable, sometimes even at diagnosis, to really detect these tumors, then really, how are we following these patients? This is really the unique challenge in the metastatic setting in patients with lobular breast cancer: we cannot rely on the imaging to tell if patients are responding to treatment. This is where liquid biopsies are really, really important, and as the field is growing up and we have better and better technologies, lobular breast cancer is going to be a field where they are going to play an important role. Dr. Rafeh Naqash: Thank you for that easy-to-understand background. The second aspect that I would like to have some context on, to help the audience understand why you did what you did, is ctDNA, tumor informed and non-informed. Could you tell us what these subtypes of liquid biopsies are and why you chose a tumor informed assay for your study? Dr. Julia Foldi: Yes, it is really important to understand these differences. As you mentioned, there are two main platforms for liquid biopsy assays, circulating tumor DNA assays. I think what is more commonly used in the metastatic setting are non-tumor informed assays, or agnostic assays. These are generally next-generation sequencing-based assays that a lot of companies offer, like Guardant, Tempus, Caris, and FoundationOne. These do not require tumor tissue; they just require a blood sample, a plasma sample, essentially. The next-generation sequencing is done on cell-free DNA that is extracted from the plasma, and it is looking for any cell-free DNA and essentially, figuring out what part of the cell-free DNA comes from the tumor is done through a bioinformatics approach. Most of these assays are panel tests for cancer-associated mutations that we know either have therapeutic significance or biologic significance. So, the results we receive from these tests generally read out specific mutations in oncogenic genes, or sometimes things like fusions where we have specific targeted drugs. Some of the newer assays can also read out tumor fraction; for example, the newest generation Guardant assay that is methylation-based, they can also quantify tumor fraction. But the disadvantage of the tumor agnostic approach is that it is a little bit less sensitive. Opposed to that, we have our tumor informed tests, and these require tumor tissue. Essentially, the tumor is sequenced; this can either be whole exome or whole genome sequencing. The newer generation assays are now using whole genome sequencing of the tumor tissue, and a personalized, patient-specific panel of alterations is essentially barcoded on that tumor tissue. This can be either structural variants or it can be mutations, but generally, these are not driver mutations, but sort of things that are present in the tumor tissue that tend to stay unchanged over time. For each particular patient, a personalized assay, if you want to call it a fingerprint or barcode, is created, and then that is what then is used to test the plasma sample. Essentially, you are looking for that specific cancer in the blood, that barcode or fingerprint in the blood. Because of this, this is a much more sensitive way of looking for ctDNA, and obviously, this detects only that particular tumor that was sequenced originally. So, it is much more sensitive and specific to that tumor that was sequenced. You can argue for both approaches in different settings. We use them in different settings because they give us different information. The tumor agnostic approach gives us mutations, which can be used to determine what the next best therapy to use is, while the tumor informed assay is more sensitive, but it is not going to give us information on therapeutic targets. However, it is quantified, and we can follow it over time to see how it changes. We think that it is going to tell us how patients respond to treatment because we see our circulating tumor DNA levels rise and fall as the cancer burden increases or decreases. We decided to use the tumor informed approach in this particular study because we were really interested in how to determine if patients are having response to treatment versus if they are going to progress on their treatment, more so than looking for specific mutations. Dr. Rafeh Naqash: When you think about these tumor informed assays and you think about barcoding the mutations on the original tumor that you try to track or follow in subsequent blood samples, plasma samples, in your experience, if you have done it in non-lobular cancers, do you think shedding from the tumor has something to do with what you capture or how much you capture? Dr. Julia Foldi: Absolutely. I think there are multiple factors that go into whether someone has detectable ctDNA or not, and that has to do with the type of cancer, the location, right, where is the metastatic site? This is something that we do not fully understand yet: what are tumors that shed more versus not? There is also clearance of ctDNA, and so how fast that clearance occurs is also something that will affect what you can detect in the blood. ctDNA is very short-lived, only has a half-life of hours, and so you can imagine that if there is little shedding and a lot of excretion, then you are not going to be detecting a lot of it. In general, in the metastatic setting, we see that we can detect ctDNA in a lot of cases, especially when patients are progressing on treatment, because we imagine their tumor burden is higher at that point. Even with the non-tumor informed assays, we detect a lot of ctDNA. Part of this study was to actually assess: what is the proportion of patients where we can have this information? Because if we are only going to be able to detect ctDNA in less than 50 percent of patients, then it is not going to be a useful method to follow them with. Because this field is new and we have not been using a lot of tumor informed assays in the metastatic setting, we did not really know what to expect when we set out to look at this. We did not know what was going to be the baseline detection rate in this patient population, so that was one of the first things that we wanted to answer. Dr. Rafeh Naqash: Excellent. Now going to this manuscript in particular, what was the research question, what was the patient population, and what was the strategy that you used to investigate some of these questions? Dr. Julia Foldi: So, we partnered with Natera, and the reason was that their Signatera tumor-informed assay was the first personalized, tumor-informed, really an MRD assay, minimal residual disease detection assay. It has been around the longest and has been pretty widely used commercially already, even though some of our data is still lacking. but we know that people are using this in the real world. We wanted to gather some real-world data specifically in lobular patients. So, we asked Natera to look at their database of commercial Signatera testing and look for patients with stage 4 lobular breast cancer. The information all comes from the submitting physicians sending in pathologic reports and clinical notes, and so they have that information from the requisitions essentially that are sent in by the ordering physician. We found 66 patients who were on first-line or close to first-line endocrine-based therapies for their metastatic lobular breast cancer and had serial collections of Signatera tests. The way we defined baseline was that the first Signatera had to be sent within three months of starting treatment. So, it is not truly baseline, but again, this is a limitation of looking at real-world data is that you are not always going to get the best time point that you need. We had over 350 samples from those 66 patients, again longitudinal ctDNA samples, and our first question was what is the baseline detection rate using this tumor informed assay? Then, most importantly, what is the concordance between changes in ctDNA and clinical response to treatment? That is defined by essentially radiologic response to treatment. Dr. Rafeh Naqash: Interesting. So, what were some of your observations in terms of ctDNA dynamics, whether baseline levels made a difference, whether subsequent levels at different time points made a difference, or subsequent levels at, let us say, cycle three made a difference? Were there any specific trends that you saw? Dr. Julia Foldi: So, first, at baseline, 95 percent of patients had detectable ctDNA, which is, I think, a really important data point because it tells us that this can be a really useful test. If we can detect it in almost all patients before they start treatment, we are going to be able to follow this longitudinally. And again, these were not true baseline samples. So, I think if we look really at baseline before starting treatment, almost all patients will have detectable ctDNA in the metastatic setting. The second important thing we saw was that disease progression correlated very well with increase in ctDNA. So, in most patients who had disease progression by imaging, we saw increase in ctDNA. Conversely, in most patients who had clinical benefit from their treatment, so they had a response or stable disease, we saw decrease in ctDNA levels. It seems that what we call molecular response based on ctDNA is tracking very nicely along with the radiographic response. So, those were really the two main observations. Again, this is a small cohort, limited by its real-world nature and the time points that ctDNA assay was sent was obviously not mandated. This is a real-world data set, and so we could not really look at specific time points like you asked about, let us say, cycle three of therapy, right? We did not have all of the right time points for all of the patients. But what we were able to do was to graph out some specific patient scenarios to illustrate how changes in ctDNA correlate with imaging response. I can talk a little bit about that. Dr. Rafeh Naqash: That was going to be my question. Did you see patients who had serial monitoring using the tumor informed ctDNA assay where the assay became positive a few months before the imaging? Did you have any of those kinds of observations? Dr. Julia Foldi: Yes, so I think this is where the field is going: are we able to use this technology to maybe detect progression before it becomes clinically apparent? Of course, there are lots of questions about: does that really matter? But it seems like, based on some of the patient scenarios that we present in the paper, that this testing can do that. So, we had a specific scenario, and this is illustrated in a figure in the paper, really showing the treatment as well as the changes in ctDNA, tumor markers, and also radiographic response. So, this particular patient was on first-line endocrine therapy and CDK4/6 inhibitor with palbociclib. Initially, she had a low-level detectable ctDNA. It became undetectable during treatment, and the patient had a couple of serial ctDNA assays that were negative, so undetectable. And then we started, after about seven months on this combination therapy, the ctDNA levels started rising. She actually had three serial ctDNA assays with increasing level of ctDNA before she even had any imaging tests. And then around the time that the ctDNA peaked, this patient had radiographic evidence of progression. There was also an NGS-based assay sent to look for specific mutations at that point. The patient was found to have an ESR1 mutation, which is very common in this patient population. She was switched to a novel oral SERD, elacestrant, and the ctDNA fell again to undetectable within the first couple months of being on elacestrant. And then a very similar thing happened: while she was on this second-line therapy, she had three serial negative ctDNA assays, and then the fourth one was positive. This was two months before the patient had a scan that showed progression again. Dr. Rafeh Naqash: And Julia, like you mentioned, this is a small sample size, limited number of patients, in this case, one patient case scenario, but provides insights into other important aspects around escalation or de-escalation of therapy where perhaps ctDNA could be used as an integral biomarker rather than an exploratory biomarker. What are some of your thoughts around that and how is the breast cancer space? I know like in GI and bladder cancer, there has been a significant uptrend in MRD assessments for therapeutic decision making. What is happening in the breast cancer space? Dr. Julia Foldi: So, super interesting. I think this is where a lot of our different fields are going. In the breast cancer space, so far, I have seen a lot of escalation attempts. It is not even necessarily in this particular setting where we are looking at dynamics of ctDNA, but in the breast cancer world, of course, we have a lot of data on resistance mutations. I mentioned ESR1 mutation in a particular patient in our study. ESR1 mutations are very common in patients with ER-positive breast cancer who are on long-term endocrine therapy, and ESR1 mutations confer resistance to aromatase inhibitors. So, that is an area that there has been a lot of interest in trying to detect ESR1 mutations earlier and switching therapy early. So, this was the basis of the SERENA-6 trial, which was presented last year at ASCO and created a lot of excitement. This was a trial where patients had non-tumor-informed NGS-based Guardant assay sent every three to six months while they were on first-line endocrine therapy with a CDK4/6 inhibitor. If they had an ESR1 mutation detected, they were randomized to either continue the same endocrine therapy or switch to an oral SERD. The trial showed that the population of patients who switched to the oral SERD did better in terms of progression-free survival than those who stayed on their original endocrine therapy. There are a lot of questions about how to use this in routine practice. Of course, it is not trivial to be sending a ctDNA assay every three to six months. The rate of detection of these mutations was relatively low in that study; again, the incidence increases in later lines of therapy. So, there are a lot of questions about whether we should be doing this in all of our first-line patients. The other question is, even the patients who stayed on their original endocrine therapy were able to stay on that for another nine months. So, there is this question of: are we switching patients too early to a new line of therapy by having this escalation approach? So, there are a lot of questions about this. As far as I know, at least in our practice, we are not using this approach just yet to escalate therapy. Time will tell how this all pans out. But I think what is even more interesting is the de-escalation question, and I think that is where tumor informed assays like Signatera and the data that our study generated can be applied. Actually, our plan is to generate some prospective data in the lobular breast cancer population, and I have an ongoing study to do that, to really be able to tease out the early ctDNA dynamics as patients first start on endocrine therapy. So, this is patients who are newly diagnosed, they are just starting on their first-line endocrine therapy, and measure, with sensitive assays, measure ctDNA dynamics in the first few months of therapy. In those patients who have a really robust response, that is where I think we can really think about de-escalation. In the patients whose ctDNA goes to undetectable after just a few weeks of therapy with just an endocrine agent, they might not even need a CDK4/6 inhibitor in their first-line treatment. So, that is an area where we are very interested in our group, and I know that other groups are looking at this too, to try to de-escalate therapy in patients who clear their ctDNA early on. Dr. Rafeh Naqash: Thank you so much. Well, lots of questions, but at the same time, progress comes through questions asked, and your project is one of those which is asking an interesting question in a rarer cancer and perhaps will lead to subsequent improvement in how we monitor these individuals and how we escalate or de-escalate therapy. Hopefully, we will get to see more of what you are working on in subsequent submissions to JCO Precision Oncology and perhaps talk more about it in a couple of years and see how the space and field is moving. Thanks again for sharing your insights. I do want to take one to two quick minutes talking about you as an investigator, Julia. If you could speak to your career pathway, your journey, the pathway to mentorship, the pathway to being a mentor, and how things have shaped for you in your personal professional growth. Dr. Julia Foldi: Sure, yeah, that is great. Thank you. So, I had a little bit of an unconventional path to clinical medicine. I actually thought I was going to be a basic scientist when I first started out. I got a PhD in Immunology right out of college and was studying not even anything cancer-related. I was studying macrophage signaling in inflammatory diseases, but I was in New York City. This was right around the time that the first checkpoint inhibitors were approved. Actually, some of my friends from my PhD program worked in Jim Allison's lab, who was the basic scientist responsible for ipilimumab. So, I got to kind of first-hand experience the excitement around bringing something from the lab into the clinic that actually changed really the course of oncology. And so, I got very excited about oncology and clinical medicine. So, I decided to kind of switch gears from there and I went back to medical school after finishing my PhD and got my MD at NYU. I knew I wanted to do oncology, so I did a research track residency and fellowship combined at Yale. I started working early on with the breast cancer team there. At the time, Lajos Pusztai was the head of translational research there at Yale, and I started working with him early in my residency and then through my fellowship. I worked on several trials with him, including a neoadjuvant checkpoint inhibitor trial in triple-negative breast cancer patients. During my last year in fellowship, I received a Conquer Cancer Young Investigator Award to study estrogen receptor heterogeneity using spatial transcriptomics in this subset of breast cancers that have intermediate estrogen receptor expression. From there, I joined the faculty at the University of Pittsburgh in 2022. So, I have been there about almost four years at this point. My interests really shifted slowly from triple-negative breast cancers towards ER-positive breast cancers. When I arrived in Pittsburgh, I started working very closely with some basic and translational researchers here who are very interested in estrogen signaling and mechanisms of resistance to endocrine therapy, and there is a large group here interested in lobular breast cancers. During my training, I was not super aware even that lobular breast cancer was a unique subtype of breast cancers, and that is, I think, changing a little bit. There is a lot more awareness in the breast cancer clinical and research community about ILC being a unique subtype, but it is not even really part of our training in fellowship, which we are trying to change. But I have become a lot more aware of this because of the research team here and through that, I have become really interested also on the clinical side. And so, we do have a Lobular Breast Cancer Research Center of Excellence here at the University of Pittsburgh and UPMC, and I am the leader on the clinical side. We have a really great team of basic and translational researchers looking at different aspects of lobular breast cancers, and some of the work that I am doing is related to this particular manuscript we discussed and the next steps, as I mentioned, a prospective study of early ctDNA dynamics in lobular patients. I also did some more clinical research work in collaboration with the NSABP looking at long-term outcomes of patients with lobular versus ductal breast cancers in some of their older trials. And so, that is, in a nutshell, a little bit about how I got here and how I became interested in ILC. Dr. Rafeh Naqash: Well, thank you for sharing those personal insights and personal journey. I am sure it will inspire other trainees, fellows, and perhaps junior faculty in trying to find their niche. The path, as you mentioned, is not always straight; it often tends to be convoluted. And then finding an area that you are interested in, taking things forward, and being persistent is often what matters. Dr. Julia Foldi: Thank you so much for having me. It was great. Dr. Rafeh Naqash: It was great chatting with you. And 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/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.
