Matthew Galsky: Thanks, Sam, and thanks for having me. The reason that we did with this study was we did a study called HCRN GU 16-257, where we gave cisplatin-based chemotherapy with immune checkpoint blockade and pursued a response guided bladder sparing strategy in patients who had a stringently defined clinical complete response. We allowed those patients the option in that study to proceed with cystectomy versus additional immune checkpoint blockade. Almost everyone chose additional immune checkpoint blockade. We showed in that study that a large proportion of those patients remain with an intact bladder without evidence of recurrence now many years later. And the update of that data with ctDNA and utDNA data was just published in PNAS a few weeks ago.
From that study, we recognized a few limitations. And I guess I'll add one nuance that 16-257 is called that because it was launched in 2016. We actually designed it in 2015. The study that I'm going to talk about is HCRN GU 20-444. You have to remember that we're talking about this in 2026, but we designed this study in 2019, and so the field is moving really, really quickly.
With that said, in our first study, we recognized a few barriers. One was that, of course we know that many patients can't get cisplatin-based chemotherapy. And the very patients who might want to pursue a strategy like this might be the exact same patients who are not candidates for cisplatin-based chemotherapy.
Sam Chang: They seem to go hand in hand, for sure.
Matthew Galsky: We wanted to make it more generalizable. That was one. The second barrier that we recognized was that... Or, I guess related to that point, around that time the studies of neoadjuvant immune checkpoint blockade as single agents were coming out, showing pathological response rates that were really comparable to what we see with cisplatin-based regimens alone. We thought maybe we can just use immunotherapy alone, particularly in this population of patients who might be a little bit more borderline for more aggressive systemic treatments. And then, of course, the other thing that we recognized as a limitation was that our clinical complete response criteria are good, but they're not perfect, and we need to develop additional biomarkers like ctDNA and utDNA.
We launched this study. This study enrolled patients who had muscle invasive bladder cancer. A little bit of nuance to this one, we enrolled patients with T2 and T3N0M0 disease. We excluded patients with T4 disease. Not that that represents a huge population of patients in our prior studies, but because there is some relationship that we've seen between the likelihood of achieving a good response and the size of the primary tumor. That might not be surprising, there's probably some biology involved there.
T2 to T3N0M0 disease, patients got two doses of pembro at the Q6 week dosing interval. And then they underwent clinical restaging, which consisted of imaging, mostly MRI. Cystoscopy with biopsies, many visible tumor or in a template if no visible tumor, and urine cytology. If all of those examinations were normal, patients were deemed to have a clinical complete response and they could pursue obser... Well, not really observation. They could pursue no definitive local therapy and receive additional pembro. And the pembro is continued in this study to complete a year of treatment. And that was based on what was happening with perioperative regimens at that time. In patients who didn't have a clinical complete response different than our initial study, to be a little bit more pragmatic, patients could have chemo radiation or cystectomy.
Sam Chang: Yeah. Either one basically in the other arm, some type of localized consolidated... You could say, they continue to have disease.
Matthew Galsky: Exactly.
Sam Chang: Great. Okay.
Matthew Galsky: We saw 43% clinical complete response rate. This is a smaller study than HCRN GU 16-257, 46 patients enrolled versus 76. And so, a smaller study, but a 43% response rate, which is actually the identical clinical complete response rate to the Gem-Cis-Nivo regimen. And I think obviously confidence intervals are somewhat wide. I'm not saying that these line up exactly with cisplatin-based chemotherapy with immune checkpoint blockade, but of course what we know from studies of neoadjuvant regimens with cisplatin-based chemotherapy with or without immune checkpoint blockade is that you do get a bump in the path CR rate, but it's certainly not a doubling. You get some bump. I think what we're seeing is probably aligned with the literature.
We only have 11 month median follow up on this study so far. And so, what we really reported was the clinical CR rate plus the ctDNA data. And the ctDNA platform that we used in this study is one that had been developed by Dan Landau's lab, was spun out in a company called CTY initially, and then acquired by Veracyte more recently. And this technology is different than Signatera. Essentially, this prioritizes breadth of sequencing rather than depth of sequencing. And what I mean by that, is that there's whole genome sequencing of both the tumor and the plasma rather than designing a Bespoke assay based on whole exome sequencing of the tumor. This is basically doing whole genome of both, and then using a bio informatics pipeline to try and match the tumor and the plasma.
Matthew Galsky: You're not limited to 16 mutations, you're limited to the whole breadth of the genome, and that arguably might be associated with better sensitivity.
What we saw was that in patients who had a clinical complete response, virtually all patients had undetectable ctDNA from day one of treatment. And so, your likelihood of having a clinical CR was markedly higher if you had baseline undetectable ctDNA. And only two patients who went on to have a clinical complete response had baseline undetectable ctDNA that converted to undetectable by their cycle two, day one assessment. One dose of six week pembro and it was gone.
In patients who didn't have a clinical complete response, the baseline ctDNA detectability was 45% of patients had detectable ctDNA at baseline. And on cycle two, day one, it was 40%. Not a whole lot of clearance in that group, much more clearance and much more baseline or ctDNA undetectability throughout in the-
Sam Chang: In the CDR?
Matthew Galsky: Yeah. As you would hope. We did report the relationship between ctDNA detectability at baseline and metastasis-free survival, and the hazard ratio was 11.2. In fact, not a single patient who had a baseline undetectable ctDNA has had metastatic recurrence. So, that's the data.
Sam Chang: And that baseline, I want to just focus on that at your last point. That baseline ctDNA, that was obviously before pembro. That was basically after the diagnosis, here we go, this is where we're at.
Matthew Galsky: Yeah.
Sam Chang: Clearly, those with invasive disease, there's going to be a cohort which we think we'll find with ctDNA that have micrometastatic disease. And then there's going to be a core that's really going to be localized. You've eliminated T4, you've focused on T2, T3. Again, I think we fine tune towards finding that population where we would obviate the need for chemo rats or cystectomy eventually. A couple things-
Matthew Galsky: Let me take it one step further, because I think it's aligned with where you're going. I think we could probably identify a population of patients who doesn't need systemic therapy. We know based on KEYNOTE-905, 9% of patients have a path CR with TRBT alone, other series that are smaller have reported higher rates, but that's a large international cohort. 9% of patients, if we could find that 9% of patients, that would probably be the simplest treatment for muscle invasive bladder cancer.
Sam Chang: Another example of beating me to the punch and showing his intelligence was... That's exactly right. Because we clearly are in an era where systemic therapy upfront's going to obviate the need for, honestly, localized therapy. Either be it chemo rads, be it cystectomy, et cetera. We're going to get to that point. I really think that. But to me, even more intriguing was, just to your point, of there's a cohort of patients we've known historically. Urologists, smaller muscle invasive tumors that just by TUR we take care of them. And it's just by luck that for some reason they haven't gone on. And we know that percentage, and it probably is under 10% or around 10%. But if we can identify those with better imaging, MRI, the ideal candidate, the repeat resection, aggressiveness, then let's not do systemic, let's not do chemo rats, et cetera. I love that. And I also think, and I want to hear what you say, I think there's a middle ground as well that, let's do a localized treat... Let's do cystectomy.
I know there's this big push against cystectomy, but let's do cystectomy and not have any systemic therapy, and we proceed. And I harken now to the CISTO data looking at patient reported outcomes for non-muscle invasive disease that have continued treatment that are recommended by guidelines to pursue surgery. Many don't. And you'll look at patient reported outcomes and physical function is actually better. I don't think we'll get rid of cystectomy, but I think we're going to get to the point where we'll do a better job of personalized medicine. Of really being able... What do you think? I'd love to hear what you think about that.
Matthew Galsky: I'm completely aligned with you, and I think there is some misinterpretation of the attempts within different camps in the field. And the goal is not to put each other out of business.
Sam Chang: Agreed.
Matthew Galsky: The goal is not to replace one modality with another. The approach is to use the minimum amount of treatment that's required to cure any single patient.
Sam Chang: No, I agree totally. Now, here's the last question, because I hate to put you on the spot. But I was wondering, and you can be totally... Take both sides, whatever. But tell me, at least logically, or perhaps with me not logically. But simply, you would think that the Veracyte type of evaluation would be a more "sensitive test". May actually pick up background noise with the evaluation. Versus something that's more focused on the 16 or 19 mutations that are most... Tell me your thoughts, pros and cons of each. I don't want you to pick a side, but tell me your thoughts about that. And should we do in both? What should we be doing?
Matthew Galsky: Yeah. I think there are clearly benefits to both. There's clearly a benefit to depth of sequencing in terms of sensitivity. You don't want to completely give up depth for breadth. But also, limiting to 16 alterations, you got to pick those 16 alterations really right. And, of course, there are platforms that try and do a little bit of both. The data that we just published from 16-257 was a collaboration with Bert Vogelstein that looked at 96 mutations per patient for the Bespoke panel, rather than 16.
And the correlation with outcomes in that study was really, really good. Is that based on different assays or based on a smaller sample size? Who knows? But I think this technology, clearly, even though there's dominant technology in clinical practice. And assays that have been commercialized obviously have an advantage in terms of momentum. This field is moving super fast and the technology's getting better and better.
Sam Chang: Yeah, for sure. And that technology being able to do exactly what you say, combine both breadth and depth. To really be able to do that just as at some point cheaply, et cetera, to get as much information as possible to get to the point of personalizing care. Matt, appreciate your time as always, your in-depth knowledge, your thoughtfulness. And honestly, always your honesty. Thanks again, and look forward to speaking to you again.
Matthew Galsky: Thank you.