Gemcitabine, Cisplatin and Immunotherapy: A Promising Mix for Advanced Bladder Cancer - Guru Sonpavde & Matthew Galsky

November 21, 2023

Guru Sonpavde and Matt Galsky delve into the CheckMate 901 study and its implications for advanced bladder cancer treatment. Dr. Sonpavde outlines the study's design, focusing on comparing gemcitabine cisplatin with and without nivolumab in cisplatin-eligible patients. He highlights the study's backdrop, noting previous trials and the potential immunogenic superiority of cisplatin. Dr. Galsky presents the results, emphasizing the study's success in meeting its co-primary endpoints of overall survival and progression-free survival. Notably, the combination arm showed a higher response rate and a significant increase in the duration of complete responses. Dr. Sonpavde discusses the study's safety profile and the impressive duration of complete responses, suggesting the combination as a new standard of care for cisplatin-eligible patients. Both experts agree on the importance of having multiple treatment options and the need for further research to identify patients who would most benefit from specific treatments.


Guru Sonpavde, MD, University of Central Florida, AdventHealth, Orlando, FL

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

Alicia Morgans, MD, MPH, Genitourinary Medical Oncologist, Medical Director of Survivorship Program at Dana-Farber Cancer Institute, Boston, MA

Read the Full Video Transcript

Alicia Morgans: Hi, I'm so excited to be here today with Dr. Guru Sonpavde and Dr. Matt Galsky, where we are going to talk about CheckMate 901 and the ESMO 2023 presentation, really trying to understand how this study and its success fit into the landscape of advanced bladder cancer with those rapid changes that we know we've seen in the last few days and months. So we'll start with you, Dr. Sonpavde. Thank you so much for joining me. Can you tell me a little bit about why the study was designed this way? What was the lay of the land when this study was designed?

Guru Sonpavde: Right, so thank you Alicia. So this is the CheckMate 901 phase three trial of which this was a component comparing gemcitabine cisplatin versus gemcitabine cisplatin plus nivolumab, of course in cisplatin eligible patients only. Now, CheckMate 901 was a complicated trial. It also had the ipi/nivo versus gemcitabine platinum comparison for PD-L1 high patients. And it also has a third comparison of gemcitabine carboplatin versus ipi/nivo in cisplatin ineligible patients. And that's not read out yet. The PD-L1 high comparison of ipi/nivo versus gem platinum has read out and was negative.

Now coming to this study, the gem/cis versus gem/cis, nivo study, it was a fairly straightforward study and the backdrop I must say is that when you look at the KEYNOTE-361 and IMvigor130 trials, which were gemcitabine platinum plus, minus either atezo or pembro. Now those trials allowed both cisplatin eligible and cisplatin ineligible patients, so they had either gem/carbo or gem/cis as the backbone. Unlike this study here, which has only cisplatin based combination chemotherapy as the backbone.

I will also point out that in both KEYNOTE-361 and IMvigor130 in the sub-analysis, there was a signal of better activity for the cisplatin gemcitabine backbone as opposed to the carbo/gem backbone for progression-free survival in both those trials actually and in the atezo trial also for survival actually. So really there was some signal that cisplatin might be a better backbone and Matt Galsky here has done a lot of work on cisplatin potentially being a more immunogenic drug, causing more immunogenic cell death. And this data that perhaps Matt can describe better than me soon in this discussion.

So really coming to the design of the study, the standard of gemcitabine cisplatin up to six cycles were allowed and in the experimental arm gem/cis up to six cycles combined with nivolumab. And once the gem/cis nivolumab combination component was completed, the nivolumab was continued in patients who were stable or responding of course up to a maximum of two years. And the co-primary endpoints were progression-free survival and overall survival. And so I guess right here we can go to Matt for the results, Alicia.

Alicia Morgans: That sounds great. So Matt would love to hear the results. Also, if you have any comments, why was this nivolumab continued for two years? Was there a thought process there before you launch into the results?

Matthew Galsky: Yeah, so the two years' sort of arbitrary in terms of what a maintenance treatment might look like and not really data-driven. We've seen maintenance strategies, treatment till progression, we've seen two years, and so this was fixed to two years.

Alicia Morgans: Great. And so what did you find?

Matthew Galsky: So the study had co-primary endpoints as noted overall survival and progression-free survival in the all comer patient population. So not PD-L1 defined patient population. And the study met its co-primary endpoints, the hazard ratio for overall survival was 0.78, and the overall survival for progression-free survival was 0.72.

Importantly, even though it wasn't a primary endpoint of this study, the response rate data was interesting. So there was a 43% response rate in the standard arm and there was a 57% response rate in the combination arm. The complete radiographic response rate doubled with gem/cis, nivo compared to gem/cis, and that was 21.7%. And of note, not only was the quantity of complete responses different, but the quality was different. So if you look at the median duration of complete responses, the median duration of complete response with gem/cis was 13.2 months and with gem/cis, nivo was 37 months, so higher response rate, higher complete response rate, different quality of complete responses, and in meeting its co-primary endpoints, the time to event endpoints.

Alicia Morgans: Very interesting. So Guru, I wonder from your perspective, what do you make of these data? I'd love to hear your thoughts not only on where they fit into the landscape, but also on the toxicity. Was there anything unexpected, anything that we should think about that's unique to this combination that is something we would not consider given a chemotherapy IO combination?

Guru Sonpavde: Right. So I think that the big highlight of this study is the duration of complete responses. So the complete response rate, as you heard, was nearly 22%, and the median duration of complete remission was 37 months approximately, So that's more than three years. That's very impressive.

Now, I will point out that historically cisplatin based chemotherapy has been known to be potentially curative in a small group of patients, and these are patients who have mostly lymph node only disease. So one of the questions that arises is in the context of EV, pembrolizumab looking very impressive overall with the 31.5 month median survival 29% complete remission rate. What do we do with gem/cis, nivo? What's the role of gem/cis, nivo? So I would argue that gem/cis, nivo has some pros in favor of it. One is the fact that the gem/cis stops at six cycles and then you continue nivolumab alone. So really the toxicity profile from that point on would be more favorable for patients. So better quality of life, I would argue after the gem/cis component is finished.

If you pick patients where you know there would be a high chance of complete remission with gem/cis, why not go with gem/cis, nivo as opposed to EV/pembro? Because with EV/pembro, as you know, the regimen was continued. Both drugs were continued until toxicity or progression. So really one of the problems with that approach of continuing till toxicity is that a lot of patients will end up having some toxicity that led to discontinuing therapy specifically neuropathy is a problem as we know with EV. So could patients coming off EV/pembro for neuropathy be compromised in terms of further delivery of subsequent therapies, especially for example, cisplatin, which is also neurotoxic?

So these are things to consider when discussing with patients. So I think that we really need to work on biomarkers that predict complete remission, so that'll help us make a more rational decision between gem/cis, nivo and EV/pembro. Of course, remember that gem/cis, nivo would only apply to cisplatin eligible patients and not cisplatin ineligible patients.

Alicia Morgans: Thank you for that. And also for reminding everyone that we did talk about the two years of nivolumab, but then treatment stops and this is something that's built into the trial which is unique. This kind of reminds me of the early conversations about you'll have your chemotherapy, then you have some time off and you get to sort of live your life and that is definitely something that can be so important for patients.

Matt, what do you think about where this fits into the landscape just in terms of tolerability schedule and all of the things that Guru was talking about with response rates considering that this would be delivered in a cisplatin eligible patient population?

Matthew Galsky: So the data with EV/pembro is spectacular, and we saw these presentations side by side, and I don't think anyone would not acknowledge that. I think the issue of course is that if we're really being thoughtful, then we have much longer follow-up. I think a year longer follow-up with gem/cis, nivo compared to the EV-302 dataset so far, that's one. And two are the points that Guru raised about the potential for a small subset of patients to be cured with cisplatin based chemotherapy.

And so I do think that we need to see what the later portions of these curves are looking like to ensure that we're not withholding a potentially curative regimen or at least a regimen that will offer a potential long treatment-free interval to a subset of patients so that we benefit the majority of patients by giving them all the same regimen.

So I'd like to see some longer follow-up from both datasets, and I'd like to see some additional work to try and refine the patients who are achieving these great responses to gem/cis, nivo. And I think we have some ideas about who those patients might be based on some of the work that's led up to these results and that just needs to be carried forward.

Alicia Morgans: I think that's fair, and I think that this is going to be the debate that we have in the field actually for the next little while given all of this data. Now just Guru, as we start to wrap up, would love to hear just your thoughts on safety tolerability of this combination. I think we are aware that things like neuropathy certainly can be something that we have to think about, hyperglycemia with EV as well as rash. There are the known side effects of chemotherapy as well as IO, which can give us effects as well and sometimes catastrophic rarely for some patients with IO but in this study, what are your thoughts and health related quality of life interestingly seem to be maintained, no real difference between the arms. What are your thoughts here?

Guru Sonpavde: Yeah, so the toxicity profile wise was quite favorable. Really, we did not see anything unexpected really what you would expect with cisplatin gemcitabine and with nivolumab, so that was reassuring. The chemotherapy, like we discussed, stops at a maximum of six cycles as the trial was designed, so really there is not the cumulative toxicity issue with the neuropathy that we would see with the EV/pembrolizumab.

In the EV/pembrolizumab trial, a median of 10 cycles was given, so at that point you do get a cumulative neuropathy. So I think that would be one big benefit of this regimen is that the cytotoxic regimen stops at a maximum of six cycles. One of the interesting things I'll point out is that more patients in the gem/cis nivo arm received all six cycles of the cis/gem chemotherapy. That was around 75% as opposed to in the chemo alone group, around 55% went up to six cycles. And historically, if you see around 50 to 60% of patients go up to six cycles of cisplatin based combination chemotherapy. So you might be targeting six, but really you cannot get there in most patients. So this suggests that with the combination of nivolumab, the responses occurred early. This is not a situation where we were converting the partial responses with gem/cis, nivo to complete responses when the maintenance nivolumab started. Really the responses occurred early. The median time to response was between two to two and a half months.
And so this higher response rate, lower rate of progressive disease as the best response, led to more chemotherapy being delivered. So around three quarters of patients getting up to six cycles of cis/gem in the cis/gem, nivo group and less chemotherapy only 55% getting to six cycles in the chemo alone group.

Alicia Morgans: Well, I have really enjoyed talking this through with both of you. Thank you for that Guru, because I think that we are all struggling with trying to pull out these nuances to try to help us inform those decisions with patients. Because in reality, they'll probably have multiple choices in this disease space to really try to get the response that they want and need. And given the fact that we want to make sure that patients have the best opportunity for them, which is a very individualized decision to start, there's going to be a lot to consider. So let's have a closing thought from both of you. Dr. Galsky, what would your message be based on this data to the viewers?

Matthew Galsky: So I would say on paper, it always seems like there needs to be one right answer for everyone. But I think all of us in practice know that there are nuances to these patient encounters that we sometimes wish we had something else because there's something small that would make them ineligible for one drug or another. So I think the more choices the better. And I think that with additional work, we might better understand these subset of patients that really have these robust responses with gem/cis, nivo.

Alicia Morgans: Great. Thank you for that. And Dr. Sonpavde, what do you think?

Guru Sonpavde: A quick word on, so there might be a question about does the javelin paradigm still apply? Why not do gem/cis followed by avelumab? Again, I will point out that the complete remission rate occurred quickly and early, and they were highly durable, 37 months median duration of complete response, 22% complete response. And really, if you reserve the PD-L1 inhibitor for after chemotherapy, you are limited to patients who had responding disease or stable disease, which you don't have with this upfront approach, and you might not get a second chance. So really I think that this approach of chemo IO in cisplatin eligible patients, I think is a legitimate new standard of care for these patients.

Alicia Morgans: Great. Well, thank you both for taking the time to talk this through with me today. I really appreciate your expertise.

Guru Sonpavde: Thank you.

Matthew Galsky: Thank you.