Phase 2 Trial: Chemotherapy Plus Immunotherapy Facilitates Bladder Sparing in Muscle-Invasive Bladder Cancer - Matthew Galsky

January 4, 2024

Ashish Kamat and Matthew Galsky discuss a phase 2 trial on organ sparing for muscle-invasive bladder cancer, published in Nature Medicine. Dr. Galsky highlights the trial's focus on systemic therapy as an alternative to cystectomy, noting its potential for curative outcomes. The study involved patients receiving gemcitabine-cisplatin plus nivolumab, followed by clinical restaging to identify clinical complete responses before surgical removal. The trial's primary endpoints were met, demonstrating that long-term bladder-intact survival is achievable in some patients treated with TURBT plus systemic therapy. Dr. Galsky emphasizes the need for more robust evidence and careful, stepwise data development before such strategies become standard, acknowledging the significant implications for patient care and treatment approaches.


Matthew Galsky, MD, Medical Oncologist, Director of GU Medical Oncology, The Tisch Cancer Institute, Mount Sinai, New York, NY

Ashish Kamat, MD, MBBS, Professor of Urology and Wayne B. Duddleston Professor of Cancer Research, University of Texas, MD Anderson Cancer Center, Houston, TX

Read the Full Video Transcript

Ashish Kamat: Hello and welcome to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat, professor of urologic oncology at MD Anderson Cancer Center, and it's a distinct pleasure to welcome again to this platform Matt Galsky, who's a good friend and expert in bladder cancer, and who is taking the time off today to spend some time with us on this very important topic, and of course, to talk about his publication in Nature Medicine that essentially deals with organ sparing for muscle-invasive bladder cancer and the data from his phase 2 trial. So Matt, thanks again for taking the time, and take it away.

Matthew Galsky: Thanks a lot for having me. So, I'm going to speak a little bit about this paper that we published recently. This was a large team effort. Many of the individuals involved are shown here, and I'll mention some specifically as we go. So, a little bit about the motivation behind this trial. So, this was a trial that we designed in 2016. And when we designed the trial, certainly, we recognized that cystectomy is a critically important pillar of treatment for muscle-invasive bladder cancer, or a potentially curative procedure.

But we recognized that not everyone can have it, not everyone should have it, and not everyone wants it. And because of those considerations, we've been thinking a lot about ways to deliver potentially curative treatment for subsets of individuals with muscle-invasive bladder cancer without the need for cystectomy. And of course, lots of individuals have thought about this over time, and there are a variety of different potential approaches to accomplish this. The approach that we've been interested in, of course, as medical oncologists, is what we can achieve with systemic therapy.

And the reason that we think about this is that we know that's a platinum-based chemotherapy — neoadjuvant chemotherapy — yields a pathological complete response in about 30 to 40% of patients with muscle-invasive bladder cancer. Of course, that's associated with better long-term outcomes when that's achieved. However, the point of administering neoadjuvant therapy, of course, is not to treat the primary tumor when the primary tumor is going to come out anyway.

The point of giving neoadjuvant chemotherapy is to eradicate micrometastatic disease that might or might not be there. However, as a byproduct, of course, we see what happens in the primary tumor and we see that a pathological complete response is achieved in a subset of patients. Unfortunately, or paradoxically, we only know that a pathological complete response has been achieved after the bladder has been removed.

And so, what if we can try and determine this before surgical removal of the bladder and allow TURBT plus systemic therapy to be advanced as a potential treatment for muscle-invasive bladder cancer? Well, of course, we're not the first ones to think about this. This has been explored over several decades. Harry Herr published on this several decades ago with the initial MVAC experience, and several others have as well.

But when we started to think about this in a comprehensive way, we recognized that there were probably some barriers to advancing this approach as a definitive treatment as one of the pillars of treatment for muscle-invasive bladder cancer. And we really outlined these five factors that we thought were preventing this approach from moving forward. The first of which is just the paucity of prospective studies. So, there are a lot of single-center experiences, and retrospective studies in the space, but very, very few prospective studies.

The second barrier that we thought was probably as important was a lack of rigorous methods to both measure and define clinical complete response. Clinical responses, being treated as a biomarker here, should be approached as rigorously as one would approach any biomarker if treatment decisions are going to be made based on this. So, we need to define it consistently, and we need to measure it consistently. The third is that there's a limited understanding of the role of delayed cystectomy in patients who initially don't have their bladders removed and experience local recurrence.

Salvage surgery, for lack of a better term, has been critical in advancing organ-sparing approaches in other cancers. And so, I think this is something that's been undermined. We probably have suboptimal systemic therapeutic regimens if we really want to advance this type of approach for a larger subset of our patients. And then, of course, an absence of biomarkers to refine this decision-making. So, with all of these barriers in mind, again, in around 2016, we designed this study.

And in this study, HCRN GU16-257, patients with muscle-invasive bladder cancer received four cycles of treatment with gemcitabine-cisplatin plus nivolumab, and then they underwent clinical restaging. And we very specifically defined what clinical restaging involved, and it involved a cystoscopy with biopsies of any visible tumor. If there wasn't any visible tumor, then biopsies of a template within the bladder, and urine cytology, and MRI of the bladder, unless MRI was contraindicated.

If all of those examinations were normal, then patients were considered to have a clinical complete response. And if they had a clinical complete response, they had the option to not undergo cystectomy and receive four more months of nivolumab as a single agent, or they could undergo cystectomy. In patients who didn't have a clinical complete response, they were recommended to undergo cystectomy. There were co-primary endpoints of the study.

The first was to define the clinical complete response rate because really, we don't know what the clinical complete response rate is with systemic treatment. It hasn't been defined consistently before. We know what the pathological CR rate might look like with neoadjuvant chemotherapy, but not really clinical complete response. And the second was to determine the positive predictive value of clinical complete response for two-year metastasis-free survival.

So, how tightly is achieving a clinical complete response linked with having long-term metastasis-free survival, or at least two-year metastasis-free survival, at least to try and show that patients who achieve a clinical complete response can safely proceed with this type of treatment approach. And sorry, I should mention that several investigators were critically involved in advancing this trial, including John Sfakianos, Sumanta Pal, and Sia Daneshmand. So, we enrolled 76 patients. 72 patients underwent clinical restaging. Four didn't for a number of reasons, which we could go into. Of the 72 patients... I'm sorry, of the 76 patients, 33 patients achieved a clinical complete response. So, the clinical complete response rate was 43%. Of those 33 patients, 32 opted for no immediate cystectomy. Only one opted for immediate cystectomy. That patient had a low-grade papillary tumor as residual disease at the time of cystectomy. And then of the 39 patients who did not achieve a clinical complete response, the vast majority underwent cystectomy.

Some underwent non-protocol therapy with radiation. These swimmers lane plots show the outcomes of the patients who achieved a clinical complete response, including the one patient who opted for immediate cystectomy. So, what you see in the swimmers lane plots, or in the dark lanes, are patients who have not had a local recurrence and are cystectomy-free, with the arrows at the end of the lane indicating patients are still alive.

You can see one of these patients, as indicated in the purple box, developed a metastatic recurrence, actually peritoneal metastases, in the absence of cancer within the bladder at that time. The light blue lanes indicate patients who had a local recurrence, as indicated in red, followed by cystectomy, as indicated in yellow. And you can see that one of those patients has had a metastatic recurrence. So two patients in that group have developed a metastatic recurrence out of the 33.

The majority though, as you can see in the Kaplan-Meier curve, remain metastasis-free. And the majority of these patients are metastasis-free with an intact bladder. So, the positive predictive value of a clinical complete response for two-year metastasis-free survival is 0.97. So, the study met its co-primary endpoints. Now, there's been a lot of interest in determining whether or not mutations in DNA damage response genes in the pre-treatment tumors can help bridge the disconnect between clinical response and pathological complete response.

So, we predefined a set of genes which, when mutated, might improve our ability beyond clinical complete response alone to predict which patients would do well with this approach. And in that panel included mutations in ERCC2, FANCC, RB1, and ATM, as defined by others, including Jonathan Rosenberg, Elie Van Allen, and Betsy Plimack.

And we were unable to show that mutations in those genes added to clinical complete response when rigorously and consistently defined in making this treatment decision in terms of potentially leaving the bladder intact after initial treatment with chemotherapy plus immune checkpoint blockade. We performed a variety of translational studies, which I'm not going to get into in detail, but simply to say that it turns out that both pre-treatment and on-treatment immune parameters were associated with the likelihood of achieving a clinical complete response to this regimen.

And for those interested in exploring that further, please see the paper for more details. In addition, in the paper, we have information on multiparametric MRI as a measure of response, and some interesting data in that regard too. So, please check out the paper. So, very briefly, how do we interpret these findings in the context of evolving data in the metastatic setting?

And I'll simply say that in the metastatic setting, combining immune checkpoint blockade with platinum-based chemotherapy has not necessarily been shown to improve outcomes in two large randomized clinical trials. However, when you look at those studies, the subset of patients who received cisplatin-based chemotherapy plus immune checkpoint blockade seemed to gain benefit with the addition of immune checkpoint blockade rather than patients who received carboplatin-based chemotherapy.

And remember, in our study, this is just cisplatin-based chemotherapy plus nivolumab. Of course, the true test of this would be a prospective study in the metastatic setting testing cisplatin-based chemotherapy plus immune checkpoint blockade like nivolumab versus cisplatin-based chemotherapy alone.

And of course, now we have the answer to that question in the metastatic setting with CheckMate 901, showing an improvement in overall survival with the same regimen that we explored in our study in the metastatic setting, gemcitabine-cisplatin plus nivolumab, and that's now published. So, in conclusion, clinical complete response is a biomarker and it should be developed rigorously as such. Long-term bladder intact survival is achievable in a subset of patients with muscle invasive bladder cancer treated with TURBT plus systemic therapy.

We think that immune checkpoint blockade adds something here, as we've seen in the metastatic setting now. Integrating novel technologies, of course, ctDNA, urine tumor DNA, will ideally refine patient selection for this approach even further. We have several ongoing studies in this space, including version 2.0 of this study, which is ongoing with immune checkpoint blockade alone, and version 3.0 will activate in 2024 with a combination regimen that I think we're all excited about.

Of course, as a community, we're going to need to define strategies to determine how we ensure that this type of approach can be compared with standard approaches to ultimately be integrated into one of the standard pillars of treatment for muscle-invasive bladder cancer.

Ashish Kamat: Thanks so much, Matt, for covering that very important paper in a short time. I really have to applaud you and the team for identifying all the key pain points, as you mentioned earlier, from the retrospective studies and addressing them in a prospective manner. I think one of the things that actually makes a lot of sense and resonates with many of us is that you didn't mandate additional systemic therapy for those with no complete response.

But, in the real world, those that don't have a response and go into a cystectomy and then have residual disease are often offered therapy. Can you comment a little bit on the real-world applicability of a trial such as this and how you think it might impact the relative outcomes?

Matthew Galsky: I think that what we're seeing from some of the randomized clinical trials integrating immune checkpoint blockade as neoadjuvant therapy either with chemotherapy or immune checkpoint blockade alone or with EV, we are seeing that patients are declining to proceed with cystectomy.

I think we really need a more robust evidence base, of course, before that becomes a standard strategy. But I think what that tells us is that patients are interested in approaches like this, and I think we owe it to the community to try and develop these strategies in a thoughtful and safe fashion. I'm not sure if I answered your question about the role of additional therapy.

Ashish Kamat: You kind of did. What I was alluding to was that, if you have the bladder preservation and you have a complete clinical response, clearly you've shown that those patients have a well-documented metastasis-free survival at the landmark time point. But, in those that don't get a complete response, were you able to glean any clues from your study as to who may or may not benefit from additional therapy?

Matthew Galsky: So, not yet, but certainly we're exploring those specimens to see if we can determine clues as to mechanisms of resistance to combination chemotherapy plus immune checkpoint blockade, as clearly that's a huge unmet need and does represent the majority of patients, unfortunately, at this time with this regimen, who don't have a complete clinical response.

Ashish Kamat: Okay, great. The other question I had for you was that, in the, and you alluded to it a little bit, but the clinical complete response was essentially an amalgamation of very rigorous biological endpoints, independent of markers. And the markers that you studied and others have studied have gone through this phase where they're believed to be changing how we look at the landscape forever, and then they go back to being research tools.

With all the work that you've done in the field, what's your current take on the role of existing markers, not the emerging novel technologies, et cetera, but the existing markers? Are they at all ready for incorporation in primetime, or should we be considering them just research tools, at least for this particular question of bladder preservation?

Matthew Galsky: Well, I think markers to predict what might happen are complicated versus markers to define what might've happened. And what I mean by that is that looking at genes that might predict benefit from treatment, we know that that calculus is so complicated in terms of what goes into a response in an individual patient. And it goes beyond having a molecular vulnerability, right? If you look at FGFR3 mutations, the majority of patients with metastatic disease with FGFR3 mutations in their tumor don't respond to an FGFR3 targeted therapy, even though you enrich those who will respond.

So, there are other things going on there, and I think that's a limitation to banking on a single gene or a set of genes to make a treatment decision such as this. Now, measuring what did happen, which is essentially what we're doing with clinical complete response in what one might be doing with ctDNA, I think that's a different story, and that speaks to our need to not only focus on pre-treatment biomarkers, which are critical because one wants to spare patients treatments that they don't ultimately benefit from.
But I think in pursuit of those biomarkers, we've actually ignored the ability to measure what's happening in an individual patient in the context of the treatment that they've received to help adapt what we do next.

Ashish Kamat: Right, right, exactly. And of course, fortunately, you didn't have too many patients that had a clinical complete response that then required a cystectomy. But I forget, from the numbers you presented, did you have sufficient numbers to be able to correlate complete response with pathologic response? Because clearly, pathologic response can only be determined once you actually remove the organ, right? Do you have those numbers?

Matthew Galsky: That's clearly the complicating aspect of this strategy, in that patients who had a clinical complete response, the vast majority didn't undergo an immediate cystectomy. Those that did had a local recurrence, so, by definition, didn't have a complete clinical response at that time. And so, it's pretty much impossible to correlate clinical and complete response with this strategy. That's certainly much better addressed in a classic neoadjuvant study.

Ashish Kamat: Right, right. Now, again, and congratulations on the results of CheckMate 901, of course. Factoring in 901 and, of course, 302 with EV Pembro, and your data, could you predict for us, and I know you talked about version 2.0 and 3.0, but looking at the landscape within the US and off-label use of drugs, et cetera, et cetera, any words of caution looking ahead and trying to predict what people might be doing off-label? Any words of caution to those listening in as to whether we can, and even if we can, whether we should be applying this paradigm to patients right now?

Matthew Galsky: So, I think we have to be really, really careful. In our purpose of pursuing a series of clinical trials and others as well, and I should have mentioned that others are very actively exploring different strategies in this space, I think we have to be quite careful in terms of doing this in a stepwise fashion and ultimately developing definitive studies.

What those studies are, I don't actually know today, but I do think that we need definitive studies before this becomes a standard treatment approach. There's a lot to gain here, but there's a lot to lose. And so, before this becomes a standard strategy, I think we really need to be thoughtful and develop the data in a stepwise manner.

Ashish Kamat: Yeah, no, well said. But I think with work ongoing, such as what you and others are doing, I think we really are going to get there much sooner in the next five to 10 years than we have in the past three decades. So, I think once again, I want to congratulate you on this excellent study, the results for our patients, and thank you for taking the time.

Matthew Galsky: Thank you.