Pembrolizumab in High-Risk, BCG Unresponsive Non-Muscle Invasive Bladder Cancer - Arjun Balar

October 24, 2020

In this conversation with Ashish Kamat, MD, Arjun Balar, MD, MBBS, discusses testing the concept of effectively treating bladder cancer without needing chemotherapy in the non-muscle-invasive disease space, specifically patients with high-risk carcinoma in situ, with or without papillary tumors that had become unresponsive to Bacillus Calmette-Guerin (BCG) based on the current Federal Drug Administration (FDA) definition. This was tested as part of this important Phase 2 single-arm study called KEYNOTE-057. 


Arjun V. Balar, MD, Associate Professor, Department of Medicine and Director of the genitourinary medical oncology program at NYU Langone’s Perlmutter Cancer Center. NYU Langone Health, New York, New York

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

Read the Full Video Transcript

Ashish Kamat:  Welcome to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat from MD Anderson Cancer Center in Houston, and it's my pleasure to welcome once again, Dr. Arjun Balar who has joined us on UroToday many times to talk about the advances in bladder cancer. Arjun is the Director of the GU Medical Oncology Program at NYU Langone Health Center. He has been active in multiple trials in the IO space, in advanced disease, and of course, in the BCG-unresponsive space as well. He's given this presentation at multiple different stages, including at the FDA when pembrolizumab was approved recently in 2020. So, again, it's always wonderful to hear from an expert on the topic. Arjun, the stage is yours.

Arjun Balar:  Thanks Ashish for inviting me once again. I feel like we've talked about this trial so much, but I think it does highlight a major and important advancement in this disease. I always like to begin with a little bit of context. Immunotherapy is no stranger in advanced bladder cancer. We've been testing checkpoint inhibitors in second-line initially and then first-line, and these two Kaplan-Meier plots highlight the activity of first-line checkpoint inhibition in advanced bladder cancer, demonstrating 23 to 29% objective response rates but really, it's all about the durability of these responses, in particular, the PRs and the CRs. It lends to what we call the tail of the curve. 

A proportion of patients who have long-term durable remissions of their cancer and perhaps effectively treating their bladder cancer without needing chemotherapy. So naturally, we wanted to test this concept in the non-muscle-invasive disease space, specifically patients with high-risk carcinoma in situ, with or without papillary tumors that had become unresponsive to BCG based on the current FDA definition. This was tested as part of this important Phase 2 single-arm study called KEYNOTE-057. Two cohorts were studied.

As I mentioned again, cohort A which was the primary focus of this trial looked at and enrolled up to 130 patients with carcinoma in situ that had become unresponsive to BCG.  And cohort B, which looked at papillary tumors alone. This study is still ongoing and eventually, we'll see data. The patients in both of these arms were treated with pembrolizumab systemically at 200 milligrams every three weeks and then underwent some routine surveillance cystoscopies and urine cytology every three months to basically monitor for efficacy.  And in the CIS population, patients have disease going into study treatment and therefore the primary efficacy endpoint that you are looking for is complete response. 

Here, the complete response rate was measured at three months, and if patients did not achieve a complete response at three months, they were discontinued for protocol therapy and counseled about cystectomy because we know cystectomy is highly curated in this patient population. So what were the baseline characteristics of the patients enrolled in this study? The efficacy population that we submitted initially to the FDA was the first 102 patients that were enrolled, the median age was 73, highly representative of the bladder cancer patient population. 

When you look at the number of prior BCG installations, you see that this is a heavily pretreated population. Again, really reflecting that this is a BCG-unresponsive disease state. If you look at the proportion of patients who had initially papillary tumors, you see that maybe about a third of patients had an additional papillary disease, including 12% of the patients who had invasive disease invading the lamina propria. We know that these are the highest risk patients and patients who really should be counseled strongly about cystectomy. Patients enrolled in this study had to clearly refuse cystectomy because we know it's curative and to be eligible for this treatment.

So, as I mentioned before, the primary efficacy endpoint that we looked at, was a complete response rate measured at three months. What we see here is that the CR rate at three months was 40%, so that means that out of the 102 patients, actually 96 that were strictly eligible per the protocol and criteria, 40% of these patients achieved a complete response at three months. And then if you look at the patients who discontinued study therapy, none of them had progressed to the muscle-invasive disease at the time of the study treatment's discontinuation. Now, ultimately some of these patients went on to get cystectomies. 

And so I'll talk about that a little bit later, but at the time of treatment discontinuation on study treatment, none had progressed to muscle-invasive disease. But, everything is about durability responses, right? So, 40% achieved responses. How long does that last? And I think this build really tells us about how to look at this. So, of the 96 efficacy evaluable patients, 39, 40% had achieved a complete response at the first disease assessment. Once again, if patients did not achieve a CR at first disease assessment, they came off study treatment.

So that means that really this population here, is 39 patients, and we call that 100%, which is the proportion that achieved a CR. And this is roughly three months after having started study treatment. What you see is that the median duration of response at this point for the population that achieved the CR was 16.2 months, and essentially about half of those, 40% had a response that lasted 12 months or longer. Now, you're looking at a median follow-up of just over two years for this population, but what you see is, if you look at all the censoring that's taking place and a flattening of the curve, very similar to what we saw in first-line and systemic checkpoint inhibition and advanced disease, is that there's clearly a plateau, a proportion of patients, roughly one out of five, that achieved long-term durable CRs. And I have several other patients who were treated in this study, and you're looking at roughly 24 months, 30 months, 33 months, two-plus years of achieving CRs. And these are the patients who have successfully avoided a cystectomy. What about clinically relevant subgroups? It's a very small study, only 96 patients. And so it's really hard to tell if there's really any subgroups that benefit more or less. I think an important question that everyone asks is, "Well, what about their PD-L1 status?"

Oddly enough, PD-L1 high tumors had a roughly lower response rate, and with a clearly overlapping confidence interval, so you really can't make much out of this, but it is somewhat interesting. This is going to be tested in a more definitive fashion as part of randomized studies including keynote 676, which is a much larger study and can really assess the prognostic and predictive value of PD-L1 expression. What's also particularly interesting is that enrollment, US versus non-US. There seems to be a separation, perhaps in terms of the complete response rates. 

I will say that one thing that is particularly interesting is that outside of the US, non-tice strains of BCG are used and it makes me wonder, are different BCG strains contributing to a subsequent immune response? This is also something that really needs to be tested in larger randomized studies. What about safety? We've always touted that systemic checkpoint blockade is really a very tolerable treatment when we compare it to chemotherapy. And that only a minority of patients have actually any significant immune-related adverse events.

But when you start treating people, some significant toxicities can develop, such as Hypophysitis, adrenal insufficiency, and Type 1 diabetes. As you can see, one patient each, developed this, these are life-altering toxicities. And so it's important to remember that when we counsel our patients, most likely nothing significant is going to happen, but some things do happen and these can be life-altering toxicities that need to be addressed in the long-term. So what about quality of life? 

This is a single-arm study, so it's very difficult to interpret quality of life on metrics. And we had several instruments that were used in this study. I will point out the CLSS score, which is really a lower urinary tract symptom score, and that's really relevant for patients with localized bladder cancer, and what you see is a suggestion, essentially, that quality of life really doesn't change all that much during the course of treatment, but there's a suggestion perhaps, that it might actually improve and perhaps urinary tract symptoms, for patients who achieve a complete response, as you can see by the solid line here, these actually might improve because maybe the CIS is getting better. 

And so some of the disease-related symptoms are perhaps getting better with treatment. Now, I wanted to talk about this, because this is something that we included as part of our FDA submission and also our updates at the annual ASCO, as well as the EUA, which is, what about the patients who didn't respond, or they responded and then progressed, or recurred? And then they went on to cystectomy? What are the consequences of delaying cystectomy in this population? So it's difficult to answer unless the patient chooses to undergo a cystectomy.

And what we know is that 36 patients at the time of this data analysis had undergone cystectomy. And what you see here is that out of those 36, only three actually were upstage to muscle-invasive disease or higher. And I've spoken to many of my colleagues in the urologic oncology field, and they would say that BCG unresponsive patients, easily 10 to 15% of these patients will be upstage at the time they are in cystectomy. To have three out of the 36, which is roughly about 8%, suggest perhaps that we are not risking upstaging of this disease, even for the patients who may not respond initially, or that respond initially and then progress, and then undergo a cystectomy that we are really not compromising the curability of cystectomy in this population. All right. So what about our conclusions? How do we put this all together? I think it's safe to say that CPIs have completely revolutionized how we think about bladder cancer therapy and how we think about this disease. What we know from this trial is that pembrolizumab induces CRs, and as a medical oncologist who always thinks about it from a systemic standpoint, when we see activity like this, it suggests a systemic activation of anti-tumor immunity.

Being a person who's not a urologist, when I think about intravesical agents versus systemic agents, it makes me wonder if CRs in this context may actually be more reliable because it's a systemically administered agent, but likely perhaps treats subepithelial disseminated disease, and this is the primary concern and the reason we counsel our patients about cystectomy in the first place.  It's a redefined multidisciplinary care, our usual model, which is sequential.  Patients come to us for chemotherapy and then we refer them back to our surgeons for a cystectomy. 

This is like home management, right? So patients need the expertise of both the urologist and the medical oncologist in parallel, to minimize the safety risks and the ongoing studies of systemic and intravesical therapies that are now being used in combination. So we have randomized studies involving intravesical BCG along with systemic checkpoint blockade. It really stresses the importance of this multidisciplinary collaboration. We always talk about biomarkers to predict response and how important they are. I can't think of a disease state where it's even any more important than localized bladder cancer in particular, because a cystectomy is highly curative in this patient population. 

And so therefore if we have a biomarker that tells us, listen, systemic checkpoint blockade is not going to benefit this patient, we need to counsel these patients about cystectomy because it is so highly curative. So that'll close and let's open to some discussion.

Ashish Kamat:  Thanks Arjun. Great points as always, and you covered a lot of data in a short time, which is wonderful. You've presented this many times and you and I were together at the ODAC. So, I want to ask you some questions that might not necessarily deal with the trial data because you presented that so well, but more tailored towards helping our audience understand the nuances when it comes to using this agent for our patients. So, what is the ideal patient in your mind that should consider using pembrolizumab versus going onto a radical cystectomy versus trying an intravesical agent, whether it's the combination chemotherapy or the gene therapy, which isn't approved yet, but a lot of people expect it might get approved.

Well, how do you consider that rationalization and explanation to the patient? 

Arjun Balar:  Yeah, so these are really important questions that really reflect that these are real-life discussions that we're having in the exam rooms. And I break it down into two categories, first is the patient who's just not a surgical candidate. Many times this is the octogenarian, has multiple medical comorbidities and cardiovascular disease, COPD, and all the things that accompany the patients who have bladder cancer, and for whom the risk of a cystectomy is substantial, including for both morbidity and mortality.

And so for those patients, really what you're doing is trying the best you can to treat their bladder cancer for as long as you can, with whatever tools you have. And so in those patients, I think using checkpoint blockade or intravesical chemotherapy, or let's say intravesical gene therapy with nadofaragene, any one of them are actually reasonable treatment choices. I wouldn't necessarily put one above the other, although even then I think an intravesical agent is less likely to lead to systemic toxicity. Let's say immune-related diabetes or pneumonitis and diarrhea.

So you might lean towards an intravesical agent first and see what you get out of it. And then, eventually use pembrolizumab. The other population is a little bit tougher, in the sense that they are good surgical candidates, but they are really looking for options that are not a cystectomy. And so, for them, it's a complicated discussion, and I've had these many times since the FDA approval. Is that listen, we have this drug, we have a 40% CR rate, but, we need to know that cystectomy is highly curative, and the setting, based on the information we have, that by doing this, we are not going to compromise your curability even if you choose to get a cystectomy later.

But you need to know for sure, that right now if you do this, this is your best choice for a curative outcome. And I think as long as patients understand that, choosing pembrolizumab versus intravesical chemotherapy I think is a reasonable choice. My opinion here is that if the patients are adamant and want to avoid a cystectomy and you're looking for something that is going to treat both the luminal disease as well as the subepithelial and micrometastatic disease, here I'd probably lean toward checkpoint blockade because, in my opinion, that at a 3 month's look at a CR, that's a little bit more of a reliable CR than something to chemotherapy.

And where a cystectomy is definitely an option for the patient, I want that CR to be as reliable as possible. And so for that reason, I might lean toward pembrolizumab in that patient population. Now if they still do not respond to pembrolizumab and they still say, "Doc, I just don't want my bladder out, then you might lean toward intravesical chemotherapy with gemcitabine and docetaxel, which I know is highly active. And then also, Instiladrin, you know, nadofaragene, which I hope will eventually be approved.

Ashish Kamat:  So, I do want to highlight some of the points that you raised, Arjun because they are very relevant for our audience.  First off, in the trial, we saw that multiple patients, the majority, in fact, elected to undergo the trial agent rather than cystectomy, even though they were candidates. And that is a very practical discussion that we have to have with the patients, is to remind them that if they are candidates for radical cystectomy and they are selecting an agent such as pembrolizumab, that that is an active decision that we and our patients are making.

The other point I want to highlight is that if we are worried about micrometastasis in our patients then, selecting a systemic agent such as pembrolizumab clearly has that advantage or we are putting something such as combination chemo or nadofaragene or even vicineum or whatever in the bladder. Along those lines, I do have a question. We are often faced with patients that have CIS in the bladder, but they also have some activity of pan-urothelial carcinoma going on with upper tract washings that are positive, et cetera, et cetera.

And again, I know it's an off-label discussion, but, have you treated those patients, and what has been your experience in patients that have bladder and upper tract CIS for example?

Arjun Balar:  Yeah. So, limited experience with that. First I'll describe the rationale and how one could potentially do it. And it comes down to two groups of people. So bladder CIS, let's say it's BCG unresponsive, and you have a unilateral upper tract disease. So let's say CIS that's extending the left ureter, gone tracking all the way up, and it's throughout, but the right is spared. There I think you might choose, and if it's BCG unresponsive, you could still treat with pembrolizumab, but I think you are probably going to monitor the bladder very closely. 

If you have persistent CIS there in the bladder, along with the other, only one urinary tract, the other is being spared ostensibly. Then I think it's reasonable then, if they are responding, then you can continue to follow this. It's a challenge to follow the upper urinary tract. I know, right? So if you have persistent cytology, you have to figure out if is it coming from the bladder? Is it coming from the ureters? Et cetera. And I think that's a technically challenging issue. But if they don't respond to pembrolizumab, I think they need a cystectomy, and you hope that, that the remaining urinary tract, that remaining unit, stays cancer-free. 

The hardest part is in bilateral disease. Now we do not see this very often, right? Ashish, to see bilateral upper tract disease, CIS diffuse, and concomitantly with bladder CIS, I can think of less than a handful of times that I've ever seen it. But in the times that you see it, what you're offering a patient, is basically a GU ectomy right? You're taking out both units, the bladder, and everything's coming out, and there, that's a big, big decision for the patient to choose, right?

So they are going to be on life long dialysis. It is highly curative, ostensibly. And so patients need to be counseled on that. And patients who are on dialysis can still have reasonable quality of life and outcomes, but there, I'm probably a bit more enthusiastic about using a checkpoint inhibitor or perhaps some of the more recently approved agents in the upper urinary tract that I might try to do, and I'm trying to push off that major surgery for as long as I possibly can, but I might be much more enthusiastic about using pembrolizumab in that setting. 

All of it is off label because we haven't really tested it because it's very difficult to monitor upper tract disease in terms of response. But I can't think of a biological reason for us not to use it. It's just more clinical and practical in terms of how we use it.

Ashish Kamat:  Right. I envy you when you say that you rarely see those patients. I see them, I feel much more often than I should. And it's really a heart-wrenching discussion to have, especially when they're in their 40s and 50s and having a systemic agent such as pembrolizumab, I think offers a real chance, obviously with the caveats of the full disclosure and discussion with the patient. One other point that I want to raise is that you mentioned this earlier when you have patients that have had multiple intravesical therapies, multiple comorbidities, and they're seeking an option as to what to do, to be honest with you, if a patient has a bladder that's so-called a cripple bladder essentially, where it's causing them a lot of problems, he or she usually will not want to try another intravesical agent.

So that's another group of patients who are considering something systemic such as pembrolizumab, which is approved. It is sometimes the first thing that you want to discuss, if they can't undergo radical cystectomy, obviously. Although sometimes patients will say, well the bladder is causing them so much misery just take it out because they don't want to actually save it. One last question for you along these lines. Obviously, the trial was done in this group of patients and it's been approved for BCG unresponsive patients, but you've done a lot of work into understanding the mechanism of action of IOs in general, not just pembrolizumab.

What is your sense from an educated sense as to the advantages of using pembrolizumab earlier, pending the studies that have to be done of course, but in patients that have a lot of CIS in the bladder, a lot of bulky tumors, and are insistent on saving their bladder, what's your sense of using it earlier in combination with either BCG or some other intravesical agent?

Arjun Balar:  So that is exciting to think about, immune synergy. I think we think about that a lot. And because we'd like to believe that, checkpoint inhibitors, which work specifically on the PD-1 pathway, that if you give it, administer it along with other agents that enhance the relevance of the PD-1 pathway, then all of a sudden you're going to get immune synergy. So what's interesting is that in prior BCG treated samples, so BCG naive and treated samples, bladder CIS, you saw an increase in PD-L1 expression after BCG.

And so in theory, if you sequentially are treating patients, let's say combination or concurrent, you could perhaps set synergy between BCG and checkpoint inhibitors in that sense. And randomized studies are addressing that. What you prove under the microscope on a tissue sample versus what you prove in a clinic in a randomized setting are often two different things, but you need to do the studies. The one agent that I'm particularly excited about, that I hope finds its way into non-muscle-invasive diseases is enfortumab.

It's a systemically administered agent. It has a toxicity profile that I think would not be attractive if it's given systemically to the non-muscle-invasive disease population. But I would like to see it be tested as an intravesical agent and to see if they can actually demonstrate activity. That would be something particularly exciting to test in combination with a checkpoint, such as pembrolizumab, in particular, because, we are seeing preliminary activity from combination studies in advanced disease, showing extraordinarily high response rates in the 70% range and waterfall plots that show deep responses, including complete responses in the 10 to 15% range. 

And so we're seeing activity with the combination of those in advanced disease that is really telling us that this is maybe the path forward in advanced bladder cancer, and perhaps we really do need to be testing it in localized disease. There, the mechanism is thought to be that monomethyl auristatin E, which is the payload with enfortumab is potentially [inaudible] in the way it causes cell death against cancer cells. And it releases tumor neoantigens into the local tumor microenvironment, which leads to enhanced T-cell priming and therefore the PD-1 pathways are [inaudible] more downstream to that. 

So I'm really excited to test that in localized bladder cancer, but I think a lot of the science that supports immune synergy, again, what we demonstrate in myogenic tissue samples ultimately needs to be tested in the clinic.

Ashish Kamat:  Yes, definitely. And of course, the target is Nectin, which is ubiquitous in all stages of bladder cancer, and intravesically, it's a clear place for it to latch onto. So yeah, I'm excited about those combination studies, and moving it into the earlier phases would really be good for our patients. This has been great Arjun, obviously, in the interest of time, we have to cut our discussion short, but I do want to thank you for taking the time from your busy schedule, stay safe and stay well.

Arjun Balar:  My pleasure. Thanks so much, Ashish.