A roundtable discussion, moderated by Peter O’Donnell, MD, discussed the advanced urothelial treatment landscape, as well as recent trial data from ESMO 2023. Dr. O’Donnell was joined by Terence Friedlander, MD; Matthew Galsky, MD; and Jonathan Rosenberg, MD.
In the next segment of the roundtable series, the panel discusses the CheckMate 901 study results and whether giving immunotherapy earlier in the treatment course is beneficial.
Dr. O’Donnell: Why don’t we transition a little bit to some of the other data sets that were presented at ESMO? Matt, I’m going to start with you for the next, which is the CheckMate 901 trial. Of course, looking at the backbone of gemcitabine/cisplatin, and could we improve upon that by comparing against gemcitabine/cisplatin plus nivolumab? Your impressions of that data set?
Dr. Galsky: The first of 3 studies that combined immune checkpoint blockade with platinum-based chemotherapy to meet the co-primary endpoints of PFS [progression-free survival] and OS [overall survival], the main difference in this data set is that it enrolled only patients who were cisplatin-eligible, so gemcitabine/cisplatin/nivolumab versus gemcitabine/cisplatin. When you look at the shapes of the curves, a lot of the action is happening in later portions of the curve. I think this is a regimen where you get some initial stability with chemotherapy, and then you see that benefit of the subset of patients who are benefiting from immune checkpoint blockade. This is way simplistic and a bunch of speculation, but I think some of the data is pointing toward patients who benefit from gemcitabine/cisplatin and who benefit from immune checkpoint blockade are likely some of the same patients. I think when you give gemcitabine/cisplatin and nivolumab, you’re making good responses in those patients into great responses. I think from an intratumoral heterogeneity standpoint, I think EV [enfortumab vedotin] and immune checkpoint blockade might actually be working on different components of the tumor, and maybe that’s why you’re getting this major boost rather than this minor boost.
Dr. O’Donnell: I’ll ask you a follow-up, because we had seen sort of a hint of this from some of the other triplet trials that had gone on, that maybe the cisplatin group would do better. Now we actually saw that in the trial that looked at just that group. Is there something different about cisplatin? Is it the patients that were enrolled?
Dr. Galsky: I think there is something different about it. I don’t think it’s purely based on the patients who were enrolled. When you look at the IMvigor130 data, even in the patients who got cisplatin versus carboplatin, taking the immune checkpoint blockade piece out of it, there is a difference in the pre-treatment tumor microenvironment, the benefit that’s achieved with cisplatin versus carboplatin, suggesting that cisplatin is doing something immunomodulatory in those settings. Then when you add immune checkpoint blockade to that, you get this extra benefit. I think there’s a difference.
Dr. O’Donnell: Terry, I’ll go to you, because you brought up earlier, this idea of getting immunotherapy earlier seems critical for you. This data set might speak to that as well, because a lot of people pointed out that the curves seem to start to diverge…
Dr. Friedlander: Later on.
Dr. O’Donnell: … at that point, right?
Dr. Friedlander: Yeah. I think going back to this question of cisplatin/carboplatin, if you just believe that giving immunotherapy earlier was beneficial, then we would’ve seen benefit in the KEYNOTE-361 study, and we would’ve seen benefit in IMvigor130. I think there is something to be said for the partner really matters. What still puzzles me a bit is that if you look in other tumor types, carboplatin-based chemotherapy, when you add a checkpoint inhibitor to that, actually does have a benefit. Like in lung cancer, that’s a pretty standard thing, to give carboplatin with a checkpoint inhibitor, where you see benefit. We don’t see that in urothelial cancer. Lung cancer, urothelial cancer, very different cancer types, different microenvironment, different mutational profiles. But it is sort of intriguing to try and pin everything on just 1 drug and say, “Oh, this is just a bad drug with immunotherapy.” That doesn’t really play out across different tumor types.
Dr. Galsky: I think some of that complexity is that it’s not just the drug; it’s the drug interacting with the pretreatment tumor microenvironment, and that’s why this has been such a mess, because those 2 pieces are hard to sort out.
Dr. Friedlander: Right. I think in urothelial cancer, we’ve constrained ourselves by putting patients with different prognoses, different baseline risk factors for doing better or worse into the same trial, and then allowing them to get different treatments in the control arms, and then trying to make sense of the results. That’s why I think CheckMate 901 was quite a smart study, because it really tried to limit the patient population to really have that nailed down such that then you could interpret the results. Clearly, the results are positive, which is great to see.
Dr. O’Donnell: Jonathan, it was great to see a positive trial, CheckMate 901, but is there a still place for gemcitabine/cisplatin then?
Dr. Rosenberg: I think the only place is going to be if we can figure out who those durable CRs [complete responses] are in advance, in a highly robust fashion. Because there is an attraction to giving just 4 or 5 months of chemotherapy, and then stopping and continuing a maintenance checkpoint. But if we can’t pick those people out of the crowd with any high degree of efficiency, I don’t see a real role in broad populations. There may be weird scenarios where you might not want to give EV/pembrolizumab. But I think there’s some tantalizing hints around DNA damage repair mutations. It’s not clear that they play a huge role in the metastatic setting compared perhaps to perioperative therapy. There may be other biomarkers that we haven’t really interrogated, but they’re going to have to be very robust if we are going to identify that subset of patients. I don’t know that we’re going to necessarily get there.
Dr. Galsky: I will just add that it’s not only the chemotherapy part in CheckMate 901 that’s a fixed duration, but actually the immune checkpoint blockade is fixed to 2 years. In the 20% of patients who had a CR, the median duration of response was 36 months. There’s a median of a year of a treatment-free interval in those patients. I mean, if that’s something that’s important, that’s I think a key piece of the CheckMate 901 data.
Dr. Friedlander: I might just add that it’s hard to figure out… Just to go back to Jonathan’s original point, someone who would be eligible for cisplatin chemotherapy, who’s ineligible for EV/pembrolizumab? You cited earlier, neuropathy as well as hyperoxemia.
Dr. Rosenberg: I didn’t say they were good for gemcitabine/cisplatin.
Dr. Friedlander: Right, which are unfortunately both the problem because of the steroids needed for cisplatin use and the neuropathy that cisplatin causes. I’ve been wrestling in my own head with who is going to get this regimen? I think from a financial perspective, gemcitabine/cisplatin and nivolumab is going to be much easier to the health system. At least here in the United States, I don’t think that that is necessarily as an important factor as it might be in other countries. It’d be interesting to see how these approvals play out across the world in terms of EV/pembrolizumab versus the gemcitabine/cisplatin/nivolumab regimen. But if I have both options in front of me, it’s very hard to know who would be eligible for one, but ineligible for the other.
Dr. O’Donnell: I’m going to go back to something Matt just brought up about this duration of the immunotherapy in the CheckMate 901, also in the EV/pembrolizumab trial, the pembrolizumab was stopped after 2 years by protocol. Now we’ve got bladder cancer patients that are living 3 years or more. Is that what we’re doing in practice? I’ll start with you, Jonathan. Do we stop our immunotherapy at 2 years just because the trials did that?
Dr. Rosenberg: I’ve generally been doing that. I’ve had a few patients flatly refuse to stop, and I haven’t fought them that hard on it, but I’ve been trying to. In really good responders, it’s not clear that it really matters to continue beyond, and there’s good data to say that many of those responses are durable. My own gestalt is that if they progress, they probably were going to progress anyway, because they often don’t respond to reinduction with immune checkpoint blockade.
Dr. O’Donnell: Quick hits from you, have you?
Dr. Galsky: I have the same experience. I find that it’s usually me trying to convince someone to stop. It’s hard for me to feel comfortable with that discussion, because then if something happens off treatment, of course you feel terrible, so it’s a hard discussion to have.
Dr. Friedlander: I think this is a strong call for biomarkers. Cell-free DNA has been used as a sort of prognostic marker for recurrence after cystectomy. Perhaps there’s a role for this in the metastatic setting, such that if patient’s cell-free DNA goes down to zero or undetectable at some point, perhaps stopping therapy would be okay. It’s not proven. We don’t have any long-term prospective data to validate that. But at least it’s an intriguing question. I’ve had patients who are past 2 years with checkpoint inhibitor, were very reluctant to stop, because this is the 1 drug that’s keeping them alive. They thought they were going to die within 9 months or 14 months of diagnosis, and here we are at 2 years or sometimes 3 years, and they really don’t want to stop.
We’ve talked about cell-free DNA, and we said, “Look, maybe if the cell-free DNA is negative multiple times in a row, maybe it’s time to stop,” and the scans look good. I’ve actually successfully stopped people who are now at least out a year from that without relapse. That’s anecdotal, so I don’t want to take too deep into that. But I think it’s a pretty intriguing biomarker for…
Dr. Rosenberg: I think we all have those anecdotes.
Dr. Galsky: I also have the anecdotes on the flip side…
Dr. Friedlander: In the reverse.
Dr. Galsky: …whereas I have a number of patients who were off treatment for other reasons. It was before ctDNA, MRD [measurable residual disease] testing became commercially available. Now testing it, you can find minimal detectable ctDNA patients who have been off treatment for a year with no evidence of disease on scan. I completely agree with you, that’s a study that needs to be done. It’s a very straightforward question.
Dr. Friedlander: Sort of like a PSA [prostate-specific antigen] from the prostate cancer world for bladder cancer.
Dr. Rosenberg: At the same time, we don’t know that those patients don’t have disease control by their immune system and that there’s a bit of tumor that’s still living in them. We don’t know what’s going to happen. I also have at least 1 person like that, they’ve been off treatment for a year-and-a-half and nothing has happened with stable. Also, we haven’t lifted the hood on those assays well enough, I think. The companies that are doing them haven’t shown us what they’re actually doing. I don’t know if they’re looking at clonal hematopoiesis that might’ve been in some of the cells in the original tumor. I have a little bit of skepticism of some of those tests still, and so to be determined as time goes on.