Read the Full Video Transcript

Sam Chang: Hello everyone. My name is Sam Chang. I'm a urologist at Vanderbilt University in Nashville, Tennessee, and we have the privilege of actually having Dr. Saum Ghodoussipour, who is an assistant professor, but that'll change soon, at Rutgers Cancer Institute, and he will be actually presenting his institution's and his group's work looking at the role of circulating tumor DNA, cell-free DNA, and evaluation of this as a biomarker, not only in serum, but in urine, in different stages and disease states of urothelial carcinoma. So first of all, thanks so much for spending some time with us and we look forward to your presentation.

Saum Ghodoussipour: Thank you so much for the opportunity and the ability to share a little bit of our work. At our institution, we've been doing a lot of work, I think as many others are, looking at this novel biomarker and specifically its role in bladder cancer. So we recently wrote a review paper, and I'm just going to go over some of the clinical implications of this test. I'm just going to briefly go over that and then hopefully we can have a quick discussion at the end. This is my outline. So essentially cell-free DNA is fragmented extracellular DNA that's released from dying cells, and is found in several different body fluids including serum, plasma and urine. Circulating tumor DNA is the portion of cell-free DNA that contains DNA released from actual tumor cells. This has gained a lot of attention in the past decade for its potential to evaluate tumor burden in real-time, becoming even more important or applicable given all the advances in genomic work and understandings of heterogeneity in bladder cancer. So I'm going to talk about some of these different roles of circulating tumor DNA.

First, we'll talk about its role in early disease, non-muscle invasive disease, how we can use it to preoperatively risk stratify patients, how it can be used to detect relapse and progression in the adjuvant setting. This is the area that I think has been getting the most amount of attention recently. And then we'll just briefly talk about some challenges and future directions. So traditionally, when we talked about circulating tumor DNA in the early studies, it was in patients with advanced disease because it was pretty logical to assume that patients with high disease burden are going to have a lot of circulating tumor DNA that can be detected. However, ctDNA can be detected in the earliest stages of bladder cancer, including patients with non-muscle invasive disease. In 2016, there was a study by Birkenkamp that looked at 12 patients with non-muscle invasive disease that had either recurred or progressed, and they found that a tumor-specific personalized assay could detect ctDNA in the plasma of 83% of these patients, and in the urine, in 87% of these patients.

Interestingly, having a higher amount of tumor DNA in the urine, looking at that, you were able to distinguish patients who progressed with their non-muscle invasive disease versus those who recurred. So essentially it's a little bit reflective of disease stage, able to predict the patients who are going to have potentially more lethal disease ultimately. Moving on from that, the same group in 2017 used droplet digital PCR to screen circulating tumor DNA for hotspot mutations and potentially targetable alterations including FGFR-3 and PIK3CA. And they were able to identify these mutations in 36% of patients. Again, non-muscle invasive disease, but they still had these mutations detectable in the plasma and the urine. Now, higher levels were associated with tumor size, grade, risk group and ultimately in those patients who did later progress. So it certainly is prognostic but potentially has a predictive role. But again, this was a smaller study. More recently, people have been using ultra-deep sequencing. Next-gen sequencing can detect variant allele frequencies, basically the presence of circulating tumor DNA, at much lower amounts.

So we can look at less than 1% of the cell-free DNA being circulating tumor DNA that can be detected with these novel techniques. An interesting study was published by Ward not too long ago, and what they did was they looked at 956 tumor samples and 884 urine samples from patients who were undergoing hematuria evaluation. 591, 165 of them ultimately ended up having bladder cancer, and then 293 patients who were on surveillance, including 29 who had active tumors at the time. And using this technique in the urine, they had a sensitivity of 87% and specificity of 85% in detecting bladder cancer. So this is a group of patients that does require cystoscopy, which aside from being invasive, can be a time burden, a cost burden for patients, and we all know how expensive bladder cancer is to care for. The sensitivity of this test was actually highest in patients with grade three tumors, went down to 86% in grade two and 70% in grade one, so there are some limitations there, but it still appeared to be a pretty useful screening test.

Now, one other study in non-muscle invasive disease that I want to highlight, I think it's interesting to, especially in relation to the work that we're doing here at the Rutgers Cancer Institute, is that intratumoral heterogeneity can be detected via ctDNA in patients with non-muscle invasive disease and it can predict response to BCG. So this was another small study of 82 patients with TAT1 papillary disease who were treated with TURBT followed by adjuvant BCG induction. They did, again, next-gen sequencing using the 861 gene panel on plasma ctDNA and the tumor DNA. They found somatic mutations in the ctDNA in these patients with non-muscle invasive disease, again, in 65% of the cohort. And again, it was higher in patients with higher stage disease, 85% T1 versus 52% in TA.

What was pretty interesting in this study is that they were able to identify intratumoral heterogeneity, which they defined as the proportion of mutations found in ctDNA that were subclonal. So the patients who had a higher than median rate of subclonal mutations, they defined as having high tumor heterogeneity and opposite if you were less than the median. And they created this molecular tumor burden index which was reflective of the heterogeneity as well as just the amount of tumor DNA that was found. And then using a cutoff score of 0.85, which was the median for this index, they saw that disease-free survival was actually much less in patients with higher heterogeneity. You can see the blue line there. The disease-free survival was 34 versus 55 months. Sorry, opposite with the colors there. And that led to a hazard ratio of 2.23 for this circulating tumor DNA detected heterogeneity to predict response.

So in summary, in the non-muscle invasive setting, it's a lot of exciting preliminary data. When I'm talking about this with residents and whatnot, they say this is very exciting, this is the future. In a lot of ways it will be the future, but it's also today. And what I mean by that is that active, ongoing clinical trials are using urinary tumor DNA as a significant component, not yet an outcome, but this is just the schema of the MoonRISe trial, which I think most of the listeners are aware of the SunRISe trial, but MoonRISe is an intravesical pretzel that's placed in the bladder that slowly releases erdafitinib, the FGFR receptor inhibitor in patients with recurrent low-grade intermediate-risk, non-muscle invasive bladder cancer. We know that FGFR mutations are higher in this cohort up to 60% depending on what reference you read, but in this study, patients are included in the study if they have a detectable FGFR receptor alteration found on urinary tumor DNA.

You can also use tumor DNA as well, but the fact that we're already at a point we're including patients for these big trials I think is just an exciting example of where we are already. Moving on from non-muscle invasive disease to invasive disease, circulating tumor DNA has been used to preoperatively risk stratify patients in some smaller retrospective studies that had promising results. First in patients who undergo radical cystectomy, the group from Mount Sinai recently put out this paper that included 112 patients with eight months of median follow-up. They collected plasma for ctDNA analysis before and after surgery, and they saw that having positive ctDNA was predictive of poor outcomes regardless of other features that we commonly look at or clinical characteristics like stage and receipt of chemotherapy. So patients who were ctDNA positive had a higher risk of having nodal metastasis, of having locally advanced disease, and they also had worse 12-month recurrence-free survival if you preoperatively had a positive circulating tumor DNA, and you see the rates there.

Now, maybe an even more interesting or important question is what is the role of circulating tumor DNA in upper tract urothelial carcinoma? One of the biggest limitations in the management of these patients is our inability to stage them as well as we do in the bladder because we can't get as good of a resection and can't do a good TURBT in the ureter. So this was a study from Moffitt that included 30 chemo-naive patients with high-grade upper tract disease who went to surgery either nephro-U or ureterectomy. They sequenced plasma with this large 152 gene panel and they saw that circulating tumor DNA was positive in 70% of patients before surgery, and that the presence of circulating tumor DNA predicted locally advanced pathology, either muscle invasive or non-organ confined disease with a sensitivity of 71% and specificity of 94%. Again, they looked at 12-month PFS and cancer-specific survival and ctDNA status was significantly correlated there as well.

Now, no ongoing trials for me to show in this space, but I think it's just an example of things to come and we can talk about that a little bit more at the end. The last component that I want to talk about right now is the ability of ctDNA to detect relapse and progression of disease. This is the part of ctDNA that I think that has been getting the most press because there's some actually randomized controlled trial data that we can use here. But first I just want to quickly talk about this study by Christiansen in 2019 that is sort of I think a very important almost seminal paper in ctDNA and outcomes of patients with bladder cancer. So they included 68 patients who underwent neoadjuvant chemotherapy and cystectomy, and they used a patient-specific 16 gene panel deep sequencing of plasma samples at the time of diagnosis before chemotherapy, at cystectomy and after cystectomy, and they saw that the presence or absence of ctDNA was strongly correlated to outcomes.

Again, not a huge randomized dataset here, but I think they have some very exciting preliminary findings specifically in the patients who had PT0 disease at the time of cystectomy, a hundred percent of them were ctDNA negative at the time of radical cystectomy. This has implications in all of the conversations that are going on now about conservative management in patients who have a clinical complete response. Circulating tumor DNA positivity after surgery identified a hundred percent of the patients who relapsed after cystectomy with a sensitivity of a hundred percent, specificity of 98%. And what was pretty interesting was that the circulating tumor DNA positivity showed up a median of 96 days before conventional imaging ever showed recurrence of disease. And you can see these Kaplan-Meier curves at the bottom, recurrence occurred in 59% of patients at 12 months if they were positive versus 0% if they were negative. There's a gigantic hazard ratio there of 129.6 or being ctDNA positive, which is obviously a big number.

The importance of this was sort of shown in IMvigor010, which we all know was the phase three trial of adjuvant atezolizumab versus observation in patients with high-risk disease after cystectomy. The original study didn't show a benefit in all patients on the original analysis, but when they did this retrospective review of their prospectively collected samples, they did see a significant role of circulating tumor DNA, so there were 581 patients who had a valuable ctDNA, 30% of these patients were ctDNA positive at the start of adjuvant therapy, and these patients had significantly worse disease-free survival compared to those who were ctDNA negative. Now, when you broke down the cohort of patients who were negative or positive, you saw that the group that was negative had no benefit with adjuvant atezolizumab, but those who were ctDNA positive did have a significant benefit with a hazard ratio of 0.58. You can see OS was also significant there as well.

Another interesting feature of this trial that I think will play a role moving forward is that the clearance of circulating tumor DNA led to improved outcomes with a big decrease in the hazard ratio of 0.26 there. That means that the patients who were ctDNA positive got atezolizumab and then went to ctDNA negative. They had the biggest benefit. So where are we going with this now today? Well, there's a couple of exciting clinical trials including InVigor 011, which is a randomized trial in patients after surgery who are ctDNA positive either at the outset or who become ctDNA positive, they're randomized to either atezolizumab or placebo looking at DFS as a primary endpoint. And then the TOMBOLA trial is a non-randomized trial in Europe that is collecting sequential ctDNA samples on these patients, and if they become positive, they go onto adjuvant atezolizumab therapy, and the results of that will be interesting. This is another trial that's using ctDNA in the adjuvant setting. This is a trial that we're accruing to here at the Rutgers Cancer Institute.

It's the MODERN trial, which is the cooperative group trial here in the US that's using nivolumab, which did show benefit in all patients in the CheckMate-274 study. But in this study, they have, I think, a very interesting and clinically relevant design where patients undergo cystectomy if they have high-risk pathology, either PT2 or node positive disease after chemo or PT3 without chemo first. They get central ctDNA testing. Those who are positive, the high-risk cohort, get randomized to FDA-approved nivolumab or their therapy is escalated to nivolumab versus relatlimab, which is a LAG-3 inhibitor that sort of enhances the immune response with checkpoint inhibitors. Versus those who are ctDNA negative, maybe a prognostically better, more favorable cohort. They get randomized to nivolumab, which did show benefit in all comers in the original trial or de-escalation of therapy to surveillance with the opportunity to go to immunotherapy if they ever convert to positive. An exciting design, and I think that probably something we'll see a little bit more of in the future.

That was just a quick summary of where we're at, the clinical implications. I think that there's a lot of future directions where we can go. This is a promising biomarker both in prognosticating outcomes after surgery, but as I showed in those original slides, you can really characterize a lot of the genome of these patients, and with the emergence of targeted therapies, it's going to be a bigger role. I think that ctDNA in contrast to past thoughts that it's really just useful in patients with high volume metastatic disease, it can be useful in all disease stages, can identify aggressive cases and see who would benefit from more aggressive therapies. There are certainly challenges remaining capturing tumor heterogeneity. It's not all going to be in the ctDNA, but we have other thoughts on where it could be. I think understanding the kinetics, that clearance that I talked about, is going to be important.

All of these studies have different assays and then there's a lot of research going on into what's the best assay, and it's a little bit beyond sort of what we think about clinically daily, but the assays can get better and I think that through the validation and prospective clinical trials, we're going to see even more. And just quickly a shout out. I think I was invited to talk about this because of some of our ongoing work and some early grants that we have, but this review paper that was put out by Sophia Bhalla, one of our medical students who's hopefully on urology soon, Rachel Passarelli, one of our residents who I'll get to apply to oncology soon, and then Antara Biswas and Subhajyoti De who run the genomics core here at the Rutgers Cancer Institute and really do impressive stuff every day. So thanks.

Sam Chang: Saum, wow, that is a fantastic review of the possibilities that we now have with both plasma and urine for both non-invasive, invasive and obviously beyond. Let's talk about kind of the last point you raised regarding how these tests are going to get better in terms of the assays that are being used. As you look at either in the urine or in the plasma, how do you think they're going to actually be used in the short term and in the long term? And what I mean by that is, are your practicing urologists now going to start collecting urine in the foreseeable near future to help determine if a cancer is non-invasive or invasive, to help determine if a patient's T1 versus T2? Tell me how you see urinary markers now are going to be used in the near future.

Saum Ghodoussipour: Yeah, it's hard to say how near it is, but I think that this will certainly play a role in the future. Biomarkers are involved in staging and risk stratification of a lot of the diseases we take care of, prostate being one example, testis being another example. So I think that this will be incorporated into risk stratification at the very least. But I don't think we're quite ready to replace cystoscopy or standard imaging yet. I think that the lead time that they've found in these studies is really thought-provoking. The fact that the ctDNA positivity happened almost a hundred days before recurrence in a lot of these patients is important, so I think that we could get to a point someday soon where we're monitoring patients with non-muscle invasive disease on either BCG or one of the novel therapies that's coming out, and we're going to have a biomarker positive disease, but no evidence of disease. And what are we going to do in those patients? Are we going to be comfortable enough to escalate to the next thing then because we're going to have a bunch of options in our toolbox, or do we need to really wait?

So I think that that's one important thing in the future in terms of the assays being better, is we just need to understand the ones that exist. Now, most of the assays that I talked about, they were tumor-specific, meaning you looked at the mutations that were present in the original tumor and you created what a lot of people call a bespoke assay. I like that word. Some of the commercial tests are considered bespoke assays. That can be very useful in increasing sensitivity and specificity, but there might be some times where you can't get tissue or you're in this situation. So there's also the possibility to do not tumor-specific assays where you just have a very accurate gene panel of mutations in bladder cancer, and I think that that's on the horizon as well. But there's a lot of non-genetic variation in bladder cancer as well that you're not going to pick up just with ctDNA, so how are we going to augment these tests? Whether it's clinical variables, hydro LVI, those things still matter.

Sam Chang: Yeah, no, I think those points are really important. Obviously, urologists would love an off-the-shelf as opposed to a bespoke type of marker, but to be able to look at the mutations within the tumor, you would think that you would have to have some tumor tissue to be the most accurate. But all right, let's switch gears to invasive cancer and the evaluation, the risk stratification, you talked about the differentiation between those patients with invasive disease prior to cystectomy, those that clearly have a benefit if they're circulating tumor DNA negative and the increased risk if they were ctDNA positive. How about afterwards? So you have a patient, and I think this is the dilemma that we're all trying to figure out that is doing a bladder-preserving type of treatment, and those patients who, after that treatment, they are ctDNA negative, everybody feels better. If they're ctDNA positive, what does that mean?

That to me is always the big question. Does it mean we have to be more aggressive with the primary? We're talking about going ahead and removing the bladder. Does it mean that we're missing something occult and something's going to happen? To me, that group of patients, I think, is the most difficult to follow. So in that group of patients who are ctDNA positive, is there any data regarding, longitudinally, if that goes down to zero, patients do better, or if it continues to rise, they have higher risk of recurrence? Tell me if there's data regarding long-term follow-up of that ctDNA in that situation.

Saum Ghodoussipour: Yeah, I think it's all data that's emerging, and I don't think we have a final answer yet. For example, the Christensen study saw that patients who converted from positive to negative, they did well. In the paper from Mount Sinai, if you were positive at any point, it predicted a worse outcome. So it's challenging to make sense of that when you're having a conversation with patients, and I think that a lot of us now have a ctDNA result to talk about even today with patients, at least I know that I do. And the way I think about it is, if a patient comes in and they've had a good response to systemic therapy and all the conventional measures that we're looking at and they want to go with bladder preservation, I still think that that should be studied in clinical trials, but patients very oftentimes reach that point outside of a clinical trial.

And I'm not a hundred percent comfortable just going off of a cystoscopy and a cytology and a biopsy. If I have a circulating tumor DNA that's negative, I'm going to feel a little bit better about it, and I talk about that with patients, but if there's positivity, I'm going to not want to mess around with that. But can I still give that patient a benefit with a major operation? I do think so. It might be a little bit aggressive of me, but I think that surgery still has a very important curative role, or even consolidative radiation at the end because of all the advances in systemic therapy that we now have. If a patient has bad pathology after chemotherapy and then surgery, that used to be a very bad outcome, but now we've got these options for adjuvant immunotherapy, and the indications for cystectomy are changing, and I think that there's going to be a lot of patients who are not going to need it and who can avoid it. But then we're going to know a little bit more clearly those who actually do need it.

And rather than having the disease progress to metastatic disease and having to go to second and third-line therapies, I think that a curative cystectomy in some ways is de-escalation therapy, but I know it's potentially an aggressive way of thinking, but that's how I feel.

Sam Chang: Well, thanks so much for your explanations, your review and obviously thought-provoking, and we're just scratching the surface and we look forward to further work from the Rutgers Cancer Institute led by you, and we very much appreciate you spending some time with us today.

Saum Ghodoussipour: My pleasure, and thank you again for the opportunity. It was an honor.