Circulating Tumor DNA - Guiding Treatment Decisions in Bladder Cancer - Lars Dyrskjøt

January 27, 2021

In this conversation with Ashish Kamat, Lars Dyrskjøt, an expert in the field of molecular markers and personalized medicine speaks about circulating tumor DNA and the potential as a tumor for guiding treatment decisions in bladder cancer. Cell-free DNA in the blood has a half-life of less than two hours, so it makes it possible to use it as a biomarker of the direct tumor burden and hence the detection of the disease also, therefore it is possible to monitor treatment and response directly. Dr. Dyrskjøt presents data from several studies of liquid biopsies that span the areas from molecular profiling and medication to the detection of residual disease, and monitoring of response mechanisms.


Lars Dyrskjøt, MSc, PhD, Professor Department of Molecular Medicine, Department of Clinical Medicine, Aarhus University Hospital/Aarhus University, Denmark

Ashish Kamat, MD, MBBS, Professor, Department of Urology, Division of Surgery, University of Texas MD Anderson Cancer Center, President, International Bladder Cancer Group (IBCG), Houston, Texas

Read the Full Video Transcript

Ashish Kamat: Welcome to UroToday's Bladder Cancer Center of Excellence. My name is Ashish Kamat, and I'm in Houston at MD Anderson Cancer Center. It's my pleasure to welcome once again, Professor Lars Dyrskjøt, who's at the Department of Molecular Medicine, and Clinical Medicine, at Aarhus University in Denmark.

Professor Dyrskjøt needs no introduction. He is an expert in the field of molecular markers and personalized medicine. And one of the areas that he and I have talked about informally many times is the discipline of circulating tumor DNA. And Professor Dyrskjøt has done so much work on this that we felt it would be very, very appropriate to have him come back and do another discussion for you, our audience, on UroToday's Bladder Cancer Center of Excellence. So, Lars, the stage is yours.

Lars Dyrskjøt: Thank you very much, Ashish, for this opportunity again, to present some of our work here. So I'm going to talk about circulating tumor DNA and discuss the potential as a tumor for guiding treatment decisions in bladder cancer.

So, what is circulating tumor DNA? Our cells shed DNA into the circulation, and this circulating DNA is termed cell-free DNA. A part of this may originate from tumor cells and contain mutations, or other somatic genomic alterations. So this is usually released from the necrotic or apoptotic cells, or from active release. This is termed circulating tumor DNA, or ctDNA, in short.

The DNA in circulation is highly fragmented, not only protected by the nucleosomes as we can see here. So it has a fragment length of about 150 base pairs. Furthermore, the mutated fraction of the DNA may constitute a very small fraction. So, therefore, there are several technical challenges to actually measure this circulating tumor DNA, which I'm not going to touch more upon today.

The mutated DNA is actually also detectable in urine samples from renal clearance, and it may also be an important resource for the detection in tumor DNA in bladder cancer and also in other cancers as well.

One important thing is that cell-free DNA in the blood has a half-life of less than two hours. So it makes it possible to use it as a biomarker of the direct tumor burden and hence the detection of the disease also. So it becomes possible to monitor treatment and response directly. We can see here, TTD and LSI actually directly correlated to tumor stage and invasiveness, as we can see here to the left, and also if we correlate it to tumor size, we can see there's a direct correlation to the variant, the new frequency that we detect, and then the tumor volume also.

What we see here is the ctDNA may have the potential to be used throughout the disease course with patients; from early screening and early diagnosis to prognostication and detection of residual disease, and finally to monitoring response and resistance, also.

So what I'll present today is data from our studies of liquid biopsies that span the areas from molecular profiling and medication to the detection of residual disease, and monitoring of response mechanisms, as well.

So what we see here is from a study we've performed, where we did a prospective study of 68 patients treated with neoadjuvant chemotherapy, and then had cystectomy afterward, and they had the standard follow-up CT scans, as we can see here. In this study, we procured blood samples from the patients at diagnosis, together with the tumor, and then during the neoadjuvant chemotherapy before cystectomy, and then at different control time points after cystectomy. And we used about 7.5 milliliters of plasma for this. We did a deep targeted sequencing to reach about 105,000 times coverage per mutation to look at.

So in this study, we actually did exome sequencing of the primary tumor, and then identified 16 clonal mutations to design a patient-specific assay, per patient. And then we used this for monitoring of the disease. And this is what we can see here. This is an example from one patient, where they have the number of the copies of a mutated DNA on the Y-axis here, and time on the X-axis, this dotted line here, represent the cystectomy time point. Then, we can see here that before cystectomy, during neoadjuvant chemotherapy regimen. We can see that the patient, the ctDNA positive, that's the red circles here, which is dropping to zero here during the chemotherapy, and then becomes zero here, just before cystectomy, and is negative again after cystectomy, and then starts to increase at this time point here.

So what we can see here is that these marks here identified the imaging scanning results, and the green marks here are negative CT scans. And this is the first positive CT scanning we see here, at this time point, actually 245 days before we see the first positive ctDNA test here.

This is an overview of all the patients that have negative ctDNA tests throughout the disease course. And all the green marks here, that's negative CT scans, and we can see all of these patients have a good outcome, this is the cystectomy time point, and that these are the samples tested before cystectomy, during the neoadjuvant chemotherapy. And these are all the different blood samples tested at the different time points, it's during the follow-up. And we can see all of these are actually ctDNA negative assays.

If we look at the patients that have a poor outcome after cystectomy, we can see that many of these patients have a ctDNA positive, actually already, many of them at the time points before the neoadjuvant chemotherapy is administered here. And we can see if we have this peripheral marker, that's a positive CT scan. Many of these instances, we see that the circulating tumor DNA is detected before we see it on the CT scans.

These are examples from patients where we see here that the ctDNA test is positive before the treatment, and then they are negative after the cystectomy. So this may actually be examples of patients where the treatment, the neoadjuvant chemotherapy is actually curing the patient, or it could be from the resection of the lymph node metastasis, for instance, at the time of cystectomy.

This is what it looks like if we make Kaplan-Meier plots, so we can see we have a very high sensitivity and specificity of this test. But the important part here is actually to see what is the positive lead time of the circulating tumor DNA, compared to the CT scanning. And in this study, we found there, a 90 day positive lead time of the ctDNA compared to the imaging. So that gives us a window of where this treatment could be initiated earlier for these patients. So we can see from this that the circulating tumor DNA is highly prognostic.

So what about the predictive value? We found here that the ctDNA dynamics are actually also very interesting if we look at this during the neoadjuvant chemotherapy. We see here is the circulating tumor DNA levels for patients that have complete pathological downstaging after the chemotherapy. And we can see for these patients that circulating tumor DNA levels drop from higher levels here to zero after the chemotherapy, indicating that the treatment is working, and it's correlating to the pathological downstaging. These are the patients where we see no pathological downstaging, and we can see that the circulating tumor DNA is not reflected in this, and we still see high levels of ctDNA after chemotherapy. Then we have a group of patients here, where we see a depletion of the circulating tumor DNA during chemo, but we actually don't see pathological downstaging.

So what is the best measure then of outcome? We try to look at this where we correlated these factors to a recurrence after cystectomy, and here we can actually see that if we look at the pathological downstaging, it's actually not significantly correlated to a recurrence after cystectomy. And this is the case when we look at circulating tumor DNA up to zero or remains positive after treatment. This is based on more of the numbers that need to be repeated in a larger series, but it sort of indicates that the ctDNA measurements may be a better handle on the disease burden during treatment also. Tumors also can be used as a real-time marker of treatment efficacy.

So these are some different schemes that we outlined that could be ways of actually using ctDNA testing in patients with muscle-invasive bladder cancer. Here, at the time of diagnosis of muscle-invasive disease, a negative ctDNA test could indicate that the new adjuvant chemotherapy would not be needed in this case, or in a positive blood sample indicate that neoadjuvant chemotherapy is needed. And then during the monitoring, during neoadjuvant chemotherapy, then, you could sort of identify patients that respond to treatment. They could, for instance, receive additional chemo cycles if they respond to treatment. And if there's no response, then a change in treatment would probably be good, probably be better for these patients. Of course, this needs to be investigated in clinical trials, and the same for the time point after cystectomy. If patients are ctDNA positive, that could indicate that they should receive adjuvant immunotherapy, for instance, or the targeted treatments.

And that's actually what we are doing now in our intervention study, we are performing here in Denmark, called TOMBOLA, treatment of metastatic bladder cancer at time of biochemical relapse following radical cystectomy. So in this study, we sample, we include patients that have muscle-invasive disease and receive a new regimen of chemotherapy, and then we sample blood samples. I look at all of these different time points, as we can see here. After cystectomy, we take frequent blood samples as well. We would do whole-exome sequencing, to identify the clonal limitations, and then develop digital PCI assays to track the circulating tumor DNA after cystectomy, and during the neoadjuvant chemotherapy.

When the patients become ctDNA positive, if they become ctDNA positive in the follow-up after cystectomy, we start immunotherapy treatment in these patients. So the primary endpoint in this study is complete response, which is defined as ctDNA depletion that should go from positive to negative, combined with normal imaging after treatment. And then we have the secondary timelines. It's all also viral in disease-specific survival, and so on. We're recruiting patients now, we're here in Denmark for this study, and we expect to finish this during 2022.

So in summary, circulating tumor DNA has many opportunities across the patient disease course to inform clinical practice. In bladder cancer, it may be particularly important to track minimal residual disease after radical cystectomy and to monitor response and resistance during a neoregimen and immunotherapy.

It has been demonstrated in multiple other cancer types that it has the clinical potential, and I'm requesting this now and that we need to demonstrate a clinical impact of the circulating tumor DNA. So can we improve overall survival by using this new molecular handle on this? Can we improve quality of life? And can we reduce the costs also for treating the patients?

Ashish Kamat: Great. Thank you so much, Lars. As always, you present your data in a very succinct, concise fashion, which is very useful for our audience. If I could ask you some basic questions because a lot of times you get these questions from folks that are reading the papers or listening to lectures. Obviously, in this study, and in many studies, the ctDNA assay is specific to the patient themselves, correct? It's based on a tumor that you select, and you sequence. Is that accurate?

Lars Dyrskjøt: Yeah, that's correct.

Ashish Kamat: And with that knowledge, is there a potential panel that you have identified, or are aware of, that could be used for tumors in general? Or is the technology such that it would always require a specific patient tumor sequencing?

Lars Dyrskjøt: So it would be ideal if we could, for instance, develop a panel that's optimal for the mutations we often see in bladder cancer patients. The problem is just that if we, and you can of course do that, but the problem is if you need to tile props for all of these different possible mutations, then you need a rather large panel. And then in order to detect circulating tumor DNA, you need to sequence this really deep as well. So, that becomes a technical challenge. It's, of course, possible. But I think in my experience, and what other people have shown also, is that if you go for smaller patient-specific panels, that may be the way to obtain this really high sensitivity and specificity needed.

Another thing is that you need this tumor handle on this because we need to know exactly what was actually present in the tumor in order to guide this analysis. If we don't do that, then there's a risk that we will make false-positive statements about, for instance, mutations arising from some other pre-malignant lesions that are not clinically relevant. Though it could arise from clonal hematopoiesis also, so that's a challenge.

Ashish Kamat: Absolutely. And I'm glad you made that point because that was going to be my next question. When you read and listen to folks talk about using ctDNA for tumor diagnosis, without a known primary tumor in the bladder, what technology do you think is most promising there, and how do you think that will pan out? Because you, of course, don't have a clear tumor to compare with or design an assay or panel. So what is your sense as to the future of that avenue of use of ctDNA?

Lars Dyrskjøt: That's a very good question, and a very interesting area to go into also. I think the biggest problem right now is to actually identify the mutations that are relevant for this. I think that we need to probably look at, not just mutations, but different molecular layers. It could be methylation markers that should be linked to, for instance, specific mutations that we know are bladder tumor-specific, and not, for instance, arising from clonal hematopoiesis, or for instance, not often seen in, for instance, pre-malignant lesions in the bladder. That's not really relevant, for instance. But I think we need to learn a lot in that direction. It could also be the DNA fragment length, for instance.

Another interesting area is actually to go into T-cell receptor clonality, as some new recent studies have shown that may be a way of detecting disease early also. But a lot of research needs to be performed in that direction before we can really avoid reporting a lot of false positives.

Ashish Kamat: Right, right. The trial that you're running, the TOMBOLA study, using ctDNA to identify which patients should get immunotherapy. It clearly has a basis in biology and makes perfect sense. And I'm sure you're excited to see some of the results that were presented at ESMO on IMvigor010, and also the CheckMate study. I think it was CheckMate 274. IMvigor010, we know, was a negative study of adjuvant therapy in patients with advanced bladder cancer. But when they looked at the ctDNA, it clearly identified those that had a poor prognosis with persistence of ctDNA, and then the clearance of the ctDNA with atezo, those patients fared better.

Now, do you think that you would be able to identify, based on your study, which patients would benefit? Or do you propose that all patients that have persistent ctDNA, and it's not based on the amount of ctDNA, should get adjuvant or additional therapy?

Lars Dyrskjøt: So you are completely right. That was exciting to see the results coming out of that because that shows us that we can actually guide who should receive the adjuvant immunotherapy. And we can see a difference in response to this also using the circulating tumor DNA. So that was very promising, but I think the ctDNA measurement, in itself, is not going to be predictive.

But I think, in the setup, we have in the TOMBOLA trial, we're going to link the circulating tumor DNA levels with the more tumor informed analysis also, where we're going to look at the exome sequencing, RNA sequencing, and other, and molecular layers in the primary tumor also, and then try to identify some predictors of response to immunotherapy. So, my hope is that we, in the future, can make a combined approach where we make a molecular profiling of the primary tumor, and then we can make an estimate of the likelihood of response to treatment based on that. And then we can use the circulating tumor DNA at the same time as a diagnostic or prognostic marker, during the treatment, and after radical cystectomy.

Ashish Kamat: Great point. And, of course, that would tie in really well with all the extensive body of work that you've done with all the other avenues, and the multi-omics, et cetera. So, that's something that we're all looking forward to hearing from you.

I'm sure I'm going to get this one question, so I'm going to ask you. Who came up with the name, TOMBOLA?

Lars Dyrskjøt: That was one of my colleagues, [inaudible]. He was looking into the different acronyms, so we thought it would be a fun way of naming it this way.

Ashish Kamat: Okay. Well, it probably came up after a couple of beers, I'm sure. No, it's a catchy name, which is very good. Lars, obviously in the interest of time we have to close, but I want to give you last, maybe 20, 30 seconds. What is the key point you would want our audience to take home, both from your study, and the field of ctDNA in general?

Lars Dyrskjøt: I think the key point is that now we actually have a very strong indication that we can use a biomarker to guide treatment. It needs to be shown in some intervention, or in randomized clinical trials, that there is a benefit to the patient also, but I think it's one of the most promising biomarkers that I've seen. And there are so many studies now focusing on trying to show the clinical benefit of this, in guiding treatment decisions. So, I think we're very close to using this as a tool in the clinics. I'm very excited about that.

Ashish Kamat: Right. Once again, Lars, thank you for taking the time and spending 30 minutes with us and the Bladder Cancer Center of Excellence. These are crazy times we live in, but hopefully, we'll be able to meet again in person soon. Until then, stay safe.

Lars Dyrskjøt: Thank you very much. You too.