The Uromigos Episode 146: Bladder Cancer Christmas Special

Jonathan Rosenberg discusses what happened in the world of bladder cancer during 2021 and what to look out for in 2022.

Episode Transcript

Brian:
Hey, welcome everyone to our second holiday special for 2021. We’re here with Jonathan Rosenberg from Memorial. And we are going to talk about bladder cancer today sort of highlights of 2021 and looking ahead to 2022. Jonathan, welcome. If you want to just quickly introduce yourself and then maybe start out with your first sort of highlight of the year, big clinically impactful data.

Jonathan Rosenberg:
Hi, Jonathan Rosenberg here. So, I am the chief of genital urinary oncology at Memorial Sloan Kettering, and I focus on bladder and urothelial cancers. And I have led many trials and continue to be active in all sorts of bladder cancer research. So, this year, I think the biggest news was the first adjuvant therapy approved in bladder cancer, the FDA approval of nivolumab based on disease free survival benefit in patients with high-risk muscle invasive disease after cystectomy.

I think this was arguably the biggest key change in the field, at least in the United States. It’s an event driven analysis, so we don’t yet have the overall survival data, but hopefully we’ll get that in the next year or so. And we’ll confirm the data at that point.

Tom:
Jonathan, I did a talk for IO ESMO. We went virtual, it was going to be a face-to-face meeting, but Switzen changed its rules at the last minute, and so. And I gave another one of my rather rambling talks. And this talk was about endpoints for adjuvant therapy. And I looked at the melanoma data, I looked obviously at the renal data.

I looked at some of the historical bladder cancer data. Chemotherapy, adjuvant chemotherapy hazard ratio 0.54, OS 0.78. And we said that was a negative trial and we didn’t think we should be doing that. That data from a PFS [inaudible 00:02:16] better than the Nivo data and it has an OS trending in the right direction with a 22% improvement. Why are we excited about adjuvant Nivo and not adjuvant chemotherapy?

Jonathan Rosenberg:
Well, I think there’s a couple of reasons for that. So, first of all, I agree with you that we’ve looked at DFS as an inferior endpoint in the past. But essentially most of the time it was the secondary endpoint not the primary endpoint, and the studies were negative as designed. It’s important to remember that when you’re thinking about the adjuvant chemotherapy studies, that they really were essentially all OS based trials.

The other point I would make is that patients who’ve had neoadjuvant chemotherapy who were roughly, I believe, half of the patients in the adjuvant [inaudible 00:03:04] study are people who would not have been eligible for adjuvant chemotherapy or even any adjuvant therapy in the prior trials. And so, for those patients who have high risk disease, despite optimal preoperative chemotherapy, there really hasn’t been any option for those people. So that represents a big unmet need. And so, I think on that reason alone, this is a big deal, but your point is well taken.

Tom:
One more point …

Jonathan Rosenberg:
It certainly comes down to study design.

Tom:
One more point, Jonathan. Negative atezolizumab trial, 0.85 for PFS, 0.85 for OS, about those numbers, I think it might have been 0.82. How do we feel that the only OS data we’ve seen so far is negative? Do we think we are going to see the same with nivolumab? And is it premature to assume OS is going to be positive in the knowledge that we already have a negative immune checkpoint to be able to study?

Jonathan Rosenberg:
What I would say to that is that the study designs were fairly different in terms of the overall structure of the study. One was an open label study where a lot of patients voted with their feet on the atezolizumab trial and did a lot of things off protocol that very possibly blunted the effect of the trial. And there were people that jumped from trial to trial, probably few, but that was … I witnessed several patients who withdrew from the atezolizumab study after finding their randomized observation.

It’s also a risk with the ongoing maturing study of AMBASSADOR, which is also open label. And so, I do worry that perhaps there was more contamination in the open label trials. And I don’t think we should look at it as it’s a fait accompli that the nivolumab trial will be negative. I do think that it’s very likely to be positive in the PD-L1 positive subset for OS, based on the DFS benefit in PD-L1 positive patients. The overall population is probably going to be a closer call, so time will tell on that.

Brian:
Jonathan, how are you using this drug in clinical practice? We haven’t seen the DFS hazard ratio for the PD-L1 negative to my knowledge. The ITT was I think 0.7, the PD-L1 was 0.5 something. So, it’s very possible the PD-L1 negative is around one …

Jonathan Rosenberg:
The benefit of the PD-L1 negatives presumably was lower although …

Brian:
Doing the math.

Jonathan Rosenberg:
Right. The benefit presumably was lower. It has to be lower.

Brian:
Has to be lower, yeah.

Jonathan Rosenberg:
Right. On the other hand …

Brian:
And I know it’s a subset and I had to get all the caveats, but in clinical practice, are you giving it to everyone?

Jonathan Rosenberg:
I’m giving it to most people. And what I would say is that PD-L1 is such a messy biomarker that I’ve found it very difficult to feel confident that my PD-L1 test is the same as Commercial Lab’s PD-L1 test is the same as a different assay that’s used. I think the PD-L1 data overall is clear as mud. I think it’s telling us that PD-L1 patients are about, might be more, PD-L1 positive tumors are more likely to respond.

But as we’ve seen throughout all the data, it doesn’t preclude response if you’re PD-L1 low. And so, making a patient selection on a biomarker that’s flawed is difficult. And given the overall population data, I haven’t been excluding patients based on PD-L1 status. [crosstalk 00:06:42] And that’s not the label …

Brian:
Let’s say it’s perfect. Let’s say you had a PD-L1 who was faithful, and you knew this patient was truly PD-L1 positive or knew he was he or she was negative, would you act on it?

Jonathan Rosenberg:
If they had prior chemotherapy, I would give it to them and if they had not had prior chemotherapy and they were a candidate for adjuvant chemotherapy, I would give them adjuvant chemotherapy, if they were PD-L1 negative. Right? If they were PD-L1 negative and platinum refractory effectively, then that would be, I would probably offer adjuvant nivolumab regardless of if the biomarker was faithful.

Tom:
Jonathan, you’re scoring about a 7 out of 10 on this answer so far. So, this is the fail part.

Brian:
That’s a C.

Tom:
Yeah. So, the ESMO guidelines committee has given a 1D recommendation, which is actually a recommendation against, weak recommendation against giving therapy. Your position is slightly different from that. It sounds like you are sort of a 1C where you are considering it. Why do you think? What’s the difference between those 2 opinions? And do you think that just the Europeans are more cautious? I suspect we want to see OS and we are not going to change our minds until we see something from OS. Do you think ultimately that’s shortsighted?

Jonathan Rosenberg:
I think it might be. As we think about the history of drug approvals for adjuvant therapy overall, in other diseases, all of them these days are based on DFS as the primary endpoint. Now you could argue that in other diseases we’ve established surrogacy and perhaps we haven’t in bladder cancer. And that at a very granular level, that’s true, I think. But why are we holding bladder cancer to such higher standards that we are not offering our patients what might be potentially curative therapy? Albeit with cost and risk of toxicity, some of which can be very severe obviously.

But I do think we’re holding our disease to a standard that we don’t hold any other disease to, breast cancer, lung cancer, et cetera. DFS is considered a valid endpoint for approval of drugs and use of drugs in the adjuvant setting.

Brian:
And I think you can make a good point that bladder cancer occurrence is generally nastier than others, right? Renal cancer recurrence can be much more indolent, et cetera. But bladder’s not good, right? Those patients do poorly, quite obviously.

Jonathan Rosenberg:
Right. I mean, so if the nivolumab overall survival data is negative, then that would belie that argument, right?

Brian:
Yeah.

Jonathan Rosenberg:
And so, if it is negative, I will rethink my thoughts on this. And I will do another podcast with you guys.

Tom:
We’re looking forward to that.

Jonathan Rosenberg:
If it’s positive …

Tom:
What we’ll do is, we’ll do an excerpt of this podcast back season …

Brian:
We don’t have the ability to do that, don’t worry.

Tom:
Well, what do you think about that now?

Jonathan Rosenberg:
So, I think, we need to see what that data shows, right? Because if the disease, maybe just giving it a relapse is going to be just as good. My suspicion is not, actually. And just as a practicing oncologist, we all know the limitations of single agent checkpoint inhibitor therapy in bladder cancer, right? When it works, it’s like holy water. And when it doesn’t work, it really just doesn’t work a lot of the time. And it’s devastating to patients.

Tom:
Let’s move to your second favorite topic, Jonathan, of the year.

Jonathan Rosenberg:
So, I think the ongoing developments in enfortumab vedotin are very, very interesting. I think the time to event data that was presented for EV Pembro in the phase 1B study is really pretty compelling in a cisplatin ineligible population with a PFS of about a year and an OS of about over 2 years. Admittedly, a small trial. Historically, we’d expect 6 to 9 months for PFS and 12 to 16 months, maybe at best, for OS. And so, these data, I think, are the data to beat in the first line cisplatin ineligible setting. And then the approval in the United States on, I would say, limited data for EV monotherapy after checkpoint without needing prior chemotherapy.

But I think it’s actually a wise approval just because there are a lot of platinum ineligible patients. And then if they do not respond to a checkpoint inhibitor in cisplatin ineligible patients, what do you? You’re left with just carboplatin. And I think EV monotherapy is a reasonable option for patients. And I think the uptake in the community has been pretty reasonable in that scenario.

Tom:
A couple quick black and white questions, Jonathan. The first, is EV the best ADC that we have in front of us at the moment?

Jonathan Rosenberg:
With caveats, I think it is. Obviously, there’s no head-to-head data between EV and sacituzumab govitecan, but the response rate is higher. The time to event endpoints, obviously, aren’t any different. But SG at the moment only has phase 2 testing. The toxicity of EV in some patients is really a concern and a problem. An SG in most patients might be a little easier to tolerate. Although, neutropenia and gastrointestinal toxicities are issues for Sacituzumab. But my own experience with the drug is that there’s more palliation with EV on a clinical basis if patients are symptomatic than I’ve observed with SG in my practice. And that’s my own comparative.

Brian:
Where do you think EV ultimately finds a home? In a year, 3 years, 5 years, where do you think we’ll be using that drug predominantly?

Tom:
In the ghost of Christmas futures.

Brian:
The ghost of Christmas future.

Jonathan Rosenberg:
I suspect it’ll be either preoperative or first line, depending on how patients present. And probably EV Pembro, if I had to put my nickel down somewhere. The perioperative therapy, I don’t know yet. But the question really will be whether or not EV Pembro in that setting can eliminate sort of so-called cancer stem cells, micromets with the same efficiency as cisplatin, which is not highly efficient, I would say, at doing that, but does it in certain patients.

And so, that question remains to be answered. But I think the first line setting, barring a big toxicity signal in the EV Pembro larger trials, I do think will probably end up as a first line standard. Either for carboplatin, may replace carboplatin. And it’s possible, it might replace cisplatin in that setting.

Tom:
I mean, the EV-302 trial is powered for both populations, it’s an ITT population. So, we’re going against both populations. And so, we either sweeten both away or neither, I guess. And I agree with you, I think that trials can be positive. And the challenge around that, around drug development, in frontline urothelial cancer right now is you need to think about what that trial’s going to be like.

There are some people in the community who feel that CTLA-4 has a role to play. And there are some people in the community including myself that think the Nivo. [crosstalk 00:14:06] A question for a friend. So, there are some people who feel that the Ipi/Nivo trial’s got a chance of being positive in the frontline setting in the biomarker positive population. What’s your feeling on that, Jonathan?

Jonathan Rosenberg:
I actually think that will likely be a positive study as well. My own experience has been with leading the efforts in CheckMate-032, which was the first study testing it in bladder cancer. And I really feel like the only patients that I have in my practice who are off therapy and being observed without recurrences years later are those who were on that trial and responded. And often had toxicity requiring stopping therapy.

Brian:
Channeling the ghost of David McDermott there, little treatment for survival.

Jonathan Rosenberg:
I know, [inaudible 00:14:59] Atkins and all of those folks. I think there’s a there, as they say. And is it enough that’s going to put the study over the line? Is the toxicity going to scare people? This is the melanoma dosing, it’s higher dose Ipi. I remember also, when I was an investigator on the trial, trying to keep track of everybody who was on corticosteroids and what doses they were on was a challenge. And that will be a challenge to people in the when, even more so than you might have in renal because of the higher dose of avelumab.

On the other hand, if we are seeing durable, maintained remissions off therapy in a 30 to 40% of patients, perhaps, maybe not that high, 30% of patients, that will be a huge advance. So, then the question sort of becomes what we are dealing with in renal, where you have Ipi/Nivo versus TKI/IO. More chronic daily toxicity with TKI/IO combinations. I think with EV Pembro, the same concept will apply. Whereas more upfront risk but maybe more late benefit with an IO/IO combination.

Brian:
Yeah.

Tom:
Jonathan, would you like to pick a third topic, or would you like me to pick one for you?

Brian:
Leading question.

Jonathan Rosenberg:
Well, I think there’s some really intriguing data, including stuff that you presented and what I presented. We’ve talked about this a little before with the TKI/IO combinations of various sorts. Particularly the FGFR3 IO combinations with the erdafitnib and pembro data and the … data.

And there’s also multiple trials percolating out there with CaboNivo with sitravatinib and nivolumab with other multi targeted TKIs that are modulating tumor microenvironment. And so, I think some of that data at first blush is not as maybe sexy as some of the ADC data, but certainly provides some reason for hope that for genomically selected subset of patients, particularly with the FGFR3 inhibitors, we’ll see some interesting data.

I’m actually looking forward to hearing the results with T-DXd over time and the results with other HER2 ADCs in bladder cancer, because I think those may also allow us to have a genomically or in this case chemically selected population patients. So, there’s some really fascinating stuff that we’re seeing the first glimmers of real activity in subsets, just like we’ve been trying to do for years.

Brian:
So, Jonathan, maybe last part focusing on big data sets next year, and you’ve started to allude to this. What phase 3s are going to read out next year? Best you can tell, public information for us.

Jonathan Rosenberg:
I actually don’t know. I have very little insight into that because I haven’t been part of any of the big phase 3 s right now. But I would say, we’re looking to hear the results of NILE, which is testing the idea of 4 drug therapy, will be certainly an expensive combination.

Brian:
And remind me, what is that trial?

Jonathan Rosenberg:
That, oh, wait, not NILE. Is it NILE, Tom?

Tom:
Yes. It’s NILE’s Durva, Treme plus chemo.

Jonathan Rosenberg:
Right, Durva, Treme, chemotherapy. So, 2 IO agents plus standard chemotherapy as first line therapy. I think, CheckMate-901 may read out, at least part of it, in the next year, year and a half. And then the non-phase 3 trial that may change practice in the United States is EV-103 Cohort K, which is EV versus EV Pembro, looking to pull out the contribution of components of pembrolizumab to the EV Pembro combination in cisplatin ineligible patients. It completed accrual this year and I’m assuming we’ll see data in the primary endpoint’s objective response rate. And this is really the confirmatory trial to support an accelerated approval for the combination in the first line setting.

Tom:
Jonathan, pembrolizumab had its label changed for a second time in this year, there was no doubt meeting. We spoke to Algen who was involved in that recently. He feels that the pembrolizumab label was probably about right now. How do you look back on this story that you’ve been involved with? Clearly you led 210, which resulted in the accelerated approval of atezolizumab. How do you look back on what we’ve learned from drug development, with all of these changes to the label? And do you think we’ve ended in the right place?

Jonathan Rosenberg:
So, I’ll start with the last question first. I do think so far, we’ve ended in the right place. As first line therapy, these drugs do leave something to be desired. And I think, KEYNOTE-361 really shows how the PD-L1 staining may not identify the population. You’re hoping it’ll identify, right? But about half the patients were positive in that trial. And I think that’s part of why it failed.

Whereas if you look at IMvigor130, which is the randomized phase 3 comparing Atezo to chemotherapy, about 20% of patients at the most, maybe even lower 10% of patients were the ones who were considered PD-L1 positive. And in that population, even though it’s a tertiary endpoint, way down on the hierarchical testing scale, there’s actually pretty robust benefit to chemotherapy in the PD-L1 positive population.

And so again if your biomarker fails, you’re not going to be able to select patients properly. So, I think getting PD-L1 out of pembro label and really reserving it to those patients who shouldn’t get chemotherapy was a good thing. We’ll see, the final analysis of IMvigor130 is still pending for OS. And if that is positive, whether it’ll change practice for the chemoimmunotherapy, I don’t know, but it may rescue atezolizumab in the first line setting for PD-L1 positive patients in the US.

Brian:
Jonathan, last question from me. Do you think we’ll ever develop a biomarker for PD-L1 monotherapy in bladder?

Jonathan Rosenberg:
I think the question is really, can we develop a multidimensional biomarker that’s robust, easy to implement and use? And I’m not sure that it’s going to be in the near future. I do think it’s possible. But PD-L1 staining alone is inadequate and we probably need some combination of TMB, maybe neo-antigens, PD-L1 and CDA cortisol infiltration. And so that’s …

Tom:
What do you make of the circulating biomarker story?

Jonathan Rosenberg:
In terms of?

Tom:
In terms of ctDNA.

Jonathan Rosenberg:
ctDNA. So, the adjuvant therapy data that you’ve put together is really very compelling from IMvigor010. And I think it’s unfortunate that it was not part of the primary analysis because now we have to redo it all, and IMvigor011 is doing that. We, I think, are going to have a way to select patients. I do wonder whether ctDNA a might be able to be used as an early non-response biomarker for first line checkpoint therapy, perhaps in patients. We have to get data to look and see.

But if you see your percentage of DNA sequencing is going up and up and up in the first cycle or to abandon treatment. So, there may be a way to actually think about how to use it in the context of checkpoint blockade as well.

Tom:
Jonathan, we did a previous Christmas podcast with Silke and Chris, which has been widely described as a disaster. And it really …

Brian:
This audio is much better.

Tom:
But during that car crash of a podcast, Silke and Chris were fantastic, of course, it was more to do with me and Brian who were struggling. But during that car crash, we did manage to get off favorite festive gifts, both given and received over the last, during your lifetime. What’s your favorite gift? And of course, we open it up to anything you’ve received.

Jonathan Rosenberg:
So, I’ve had some quite spectacularly wonderful gifts from patients at the holidays, including a Hanukkah diorama filled with chocolate Hanukkah gelt, which was a 3-dimensional, like 2 feet high, 3 feet wide, foam and hot glue gun masterpiece. And it really was a shame to open it up, but I still have it tucked away in my office. So, that really was the walking out …

Brian:
That’s like a life size gift. I’m not sure that …

Tom:
That’s a proper gift.

Jonathan Rosenberg:
Right. It really was. And that patient was someone who’s done very, very well on immunotherapy and was grateful for everything that the field done. And so, I was the beneficiary of that gift. And things like that really make your day.

Brian:
For sure.

Tom:
Jonathan, this has been terrific. Thank you very much for your time. Have a terrific next couple of weeks and a lovely Christmas. And thanks as always for joining us. And we’ll see you in the new year.

Brian:
Thanks, Jonathan.

Jonathan Rosenberg:
My pleasure. Thanks so much.