Matthew Zibelman: Yeah, absolutely. Ashish, thanks so much for having me. I really appreciate it. So I'm going to talk to you today about our presentation that we did recently at GU ASCO, so I'll move forward. So this is a post-hoc analysis of a prior study that I did in conjunction with Alex Kutikov, one of the urologists here at Fox Chase Cancer Center. And really the idea was very simple. We know that for patients with localized bladder cancer, that clinical staging is inadequate to predict pathologic staging ultimately, but that had never really been looked at well prospectively. And so, the urologists had really a very simple premise. They thought that if they did a very systematic cystoscopy inside the bladder prior to patients going for planned cystectomy, that they would be able to better predict patients who were ultimately going to have residual disease. And so, we designed a study to look at this prospectively.
We included patients both with muscle-invasive bladder cancer and non-muscle-invasive bladder cancer, really anyone going for planned cystectomy. For the muscle-invasive patients, they could have had prior neoadjuvant chemotherapy or not. And basically, the urologist at the time of surgery in the operating room did a cystoscopy with very focused biopsies for areas that had scar or any abnormality. And then, we looked at what they saw, what the pathology from that predicted, and how it compared to ultimately the final pathology. And this was published in Journal of Urology in 2021. The primary endpoint was the negative predictive value of what we determined our S-E-E or SEE, because we were seeing what happened with this systematic endoscopic evaluation. And basically, I'll move to the next slide. This is just a schema of how we did it, patients with muscle-invasive or non-muscle-invasive bladder cancer with or without neoadjuvant chemotherapy went to surgery, they did this SEE, and then we compared it to the final pathology. And the key takeaway from the trial is this slide here with this graph. Basically, patients who we saw nothing, they were SEE T0, about a quarter of them had muscle-invasive bladder cancer, so T2 or higher disease.
So even with a very systematic approach, we missed a quarter of muscle-invasive bladder cancers. And so to us, this was significant, clearly proving that clinical staging was inadequate. And so what I'm presenting now is a post-hoc analysis of this by taking some blood that we had collected at the time of surgery, the day of the surgery and SEE to look at circulating tumor DNA to see if adding that helped us predict what was going to be inside the bladder. So this is what I presented at GU ASCO. I have some of the statistical details, but I'll just cut to the point. Basically, we used a ctDNA assay from a company, GRAIL. One key about this is this is a tumor-naive assay as opposed to the approaches being used now, which require tumor to be able to predict the ctDNA. This is done the different methylation-based approach, so we don't need tumor potentially cutting down the time to get results. And really what we wanted to see was if adding ctDNA could help us in our prediction. On the right here, I have the breakdown of the patients who we were able to analyze retrospectively. It was 54 patients.
I have them split as patients who had disease at surgery versus those that were pT0. You see most of the patients had muscle-invasive bladder cancer, but about 25 to 30% had non-muscle-invasive. You see the majority had received neoadjuvant chemotherapy where applicable. The take-home of this analysis is that when we added them together, so we added patients who had both the SEE, we looked inside the bladder and didn't see anything, and ctDNA was negative, our negative predictive value for finding muscle-invasive disease was way better. It went from 48% up to 83%. And so, this tool is getting closer using these combined, we think, to better help predict who is going to have residual disease inside the bladder. And as we move forward to potential bladder-sparing approaches, we think these combination-type markers will help us better predict who may be safely able to consider a bladder preservation type approach. The other thing we looked at and I presented was that ctDNA was a good prognostic marker. This isn't necessarily news, but similar to how we consider pT0 as a prognostic marker, both for overall survival as well as recurrence-free survival. And so, if you just look at these curves at the top is ctDNA, the bottom is the pT0 rate, and you see there's a clear separation if patients are ctDNA negative, relatively similar to the separation we see from pT0 patients versus those who have residual tumor at the time of surgery.
And so, this ctDNA tumor-naive tool seems to be prognostic, similar to pT0 and also to the other ctDNA approaches. So in summary, in this post-hoc analysis of our prospective study, combining this SEE, looking inside the bladder with ctDNA, improved our negative predictive value for deciding who might have muscle-invasive cancer at the time of surgery. Negative ctDNA is a good prognostic marker similar to pT0 in prognosticating overall survival and progression-free survival. And we think that these tumor-naive ctDNA approaches may be applicable moving forward, either along with or as good as the tumor-informed approaches, especially since the time to result may be much quicker moving forward. And so, really my take-home message is just that we know ctDNA is an exciting new part of bladder cancer. We think it likely will have a role in making decisions at the time of surgery for patients with muscle-invasive disease, and will be really a key in assessing future bladder preservation approaches for these patients. So I'll end there.
Ashish Kamat: Thanks so much, Matt. Very, very nice presentation, and I'm glad you went through all the little details and nuances. This whole concept of clinical CR, pT0, ctDNA, biopsies, really, you guys have been some of the leaders in actually addressing that very systematically. I was privileged enough to lead the IBCG Milan consensus effort that we published recently in JCO, and I did that with Andrea Necchi, Matt Galsky, of course Betsy Plimack from your institution was an integral part of that. And in there, we bring up the same points that you brought up. We have to be very careful. We can't just do endoscopic evaluation. I would love to, as a urologic oncologist, say we're great, but we're not. You need markers such as ctDNA, et cetera, et cetera. And even then, we can't get to that 95, 99. We can't get quite there. So a question to you, and some of it may be from work that you've done, some of it maybe from just your knowledge of the field, how do you factor in imaging biopsies, ctDNA when you are considering this global picture of trying to predict who has a clinical complete response that is actually going to translate in a path response, which means we shouldn't then maybe not do a cystectomy. We don't know that yet, but what's your sense?
Matthew Zibelman: Yeah, so I think one thing definitely missing from this, an earlier analysis was an imaging component. Patients generally did have a CT scan looking for metastatic disease, but we did not do an MRI to evaluate that. The reasons we didn't include that in that initial study was, one, this was unfunded, so insurance didn't always pay for it, particularly for non-muscle-invasive patients. And two, patients who'd had recent TURBTs or analysis, we were concerned that that might muddy the water somewhat. But I do think imaging can hopefully be helpful and layering that on may prove to be an additional tool. And I think if you look at Matt Galsky's initial bladder preservation trial, MRI did seem to be helpful in those patients. So we are incorporating that into our approaches moving forward and really looking at what do we see inside the bladder? What does imaging with CT and MRI show? And then, I think the next step is adding ctDNA and what that tells us. And I think, and I suspect that the final piece of this may be urine DNA. And we did collect that as part of this trial. I did not report it in this analysis, but hoping to include that in a future one. But that I think is maybe the last piece to really help us understand what's inside the muscle and in the bloodstream, and that's what ctDNA helps tell us. And then, is there a better way to know what's happening in the lining of the bladder and inside the bladder? And can urine add that final piece so that we can move to 90 to 95% predictive capacity and really allow patients to choose bladder preservation approaches confidently or at least be able to study it better in future trials.
Ashish Kamat: Yeah, you preempted a question I was going to ask you, which is great. I was just going to say, as part of the criteria that we published, and again, it's in the consensus document from the IBCG Milan effort is essentially exactly what you said. We need imaging. If you have MRI, which is I think going to be very useful, use it. If not, CT scan because many parts of the world, we don't have CT scan and then ctDNA biopsies, all of that. The question of the urinary markers, I was going to ask you again, and I'll just throw out first a comment and then hear your response. I think ctDNA clearly is very useful from a systemic perspective. When we do utDNA and others have looked at this as well, you're recognizing a field effect, recognizing changes that have occurred or are occurring in the bladder that are part of the pathogenesis. I think we shouldn't forget the utility of the UroVysion FISH, which looks at 3, 7, 17, 9p deletion, it's been around forever, has a very high predictive power and also regular cytology. So in your studies that you're doing, are you looking at all three? Are you looking at utDNA? What's your focus?
Matthew Zibelman: I mean, in our studies that we've done here, we've done usually cytology and utDNA. We have not used the FISH at this point. And I do worry about some of that field effect and the false-positive rate of urinary DNA may be a little bit problematic and we have to refine that a little bit. I do worry there's a field effect and even premalignant clones in the bladder that may yield ctDNA that looks positive, but doesn't actually reflect malignant disease. Also, the upper tracts. Some of these patients may have a cold upper tract disease that we're not aware of that's shedding DNA into the urine and we're detecting that. In fact, some of these studies, I'm curious of some of these patients with pT0 but positive urine DNA, are they going to show up to have an upper tract tumor later down the line that we maybe accidentally discovered early? So I definitely think we need to continue to refine that approach, but I'm cautiously optimistic that we'll get there and feel confident with those assays eventually.
Ashish Kamat: Great points. Matt, thank you so much for taking the time and congratulations.
Matthew Zibelman: Thank you so much. Thanks for having me.