Read the Full Video Transcript

Sam Chang: Hi, my name is Sam Chang, and I'm quite fortunate to be joined by a true superstar in urologic oncology. I've known Dr. Eugene Pietzak for years as he was a resident at the University of Pennsylvania, then a fellow at Memorial Sloan Kettering, and now one of their leading clinicians focused on urothelial carcinoma, both upper tract as well as bladder cancer.

Eugene was actually tapped by the SUO to present a session on BCG-unresponsive disease and gave a wonderful overview on options that are available currently and in the near future for BCG-unresponsive disease. So Gene, thanks again for giving that wonderful presentation, and we look forward to your overview of what you presented.

Eugene Pietzak: Thank you, Sam. Thank you for the invitation to join you for UroToday to discuss the panel session that I participated in, focused on what are our currently available options for BCG-unresponsive disease. These are my disclosures. They're relevant to drug development as well as consulting for some of these companies.

So in many ways, this is a very exciting time for urologists. Historically, there haven't really been that many treatment options available for patients who have experienced BCG failure. And now, we're like kids in candy stores, where there's a ton of different options. But it's really hard to understand the relative risk-benefit trade-offs between them because of the way that they were approved in single-arm, non-randomized trials.

And so the presentation that I gave is focused on what is currently available and what will be available in the very near future. It's also important to mention that in addition to these FDA-approved drugs, there's also several treatments that we give off-label, intravesical chemotherapies such as gemcitabine with or without docetaxel. And there's mitomycin. Technically, there's also valrubicin, although I personally have never treated a patient with that, and I don't actually know if it's any further available clinically.

But throughout this meeting and pretty much every meeting that's occurred in the last couple of years, we see these comparative slides where even though these were single-arm, non-randomized trials, we know that we should not do cross-comparisons, but we do them anyway. And I actually think that that's inappropriate and it can be misleading.

And the reason that is, I think there's some differences in the aspects of these trials that are important that make it really difficult for us to do any comparative effectiveness type of analysis across trials. And that is, there's different surgeons participating in them, and we know that there's significant variation in the quality of TURBTs that are performed.

There's also what could be termed the treatment center effect, which is, in many ways, the selection bias—whether one surgeon is favoring more cystectomy for such patients versus clinical trials versus off-label chemotherapies. We all see different patients as well in different referral patterns and where those patients were initially treated.

And so, I think it was James McKiernan at Columbia who came up with the term of urologist-unresponsive disease. But some patients may not necessarily have biologically BCG-resistant disease; they just may have been mismanaged initially. And I think, unfortunately, a lot of us see those folks show up to our clinic. And so those patients may respond differently to treatment.

And even though the trials are typically written where there should be consistency in using, say, blue light or narrowband imaging across both your TURBT versus your follow-up, that's not reported necessarily. And we don't know necessarily the consistency and how often patients who had a blue light TURBT are necessarily getting a blue light flexible cystoscopy for surveillance. And that could certainly affect the recurrence rates.

There's also what I'm coming up with—the term pathologist-dependent disease. We know that there's a lot of variation between pathologists in terms of calling something carcinoma in situ versus dysplasia, and even high-grade versus low-grade. And we deal with mixed tumors sometimes. And when your treatment failure event is high-grade recurrence, that could certainly alter the readout of a single-arm study.

And we also know, obviously, there's a lot of variation amongst the biology of the tumors themselves. And so, I do feel that what doesn't get discussed often is that there's a lot of underrepresentation of patients with high-grade T1 with CIS in these trials; they're typically only about 10% of these patients. And there's different treatment histories that patients have experienced.

And some patients that may have been previously treated with another salvage regimen, their tumor microenvironment may be different. There may be some cross-resistance that's developing, especially if patients are getting treated sequentially with multiple therapies. And I also highlight that we actually don't know the natural history of contemporary CIS.

BCG-unresponsive criteria were based off of expert opinion and not necessarily robust data. And we don't know necessarily if it's just a focal amount of CIS that's associated with the tumor versus multifocal or diffuse CIS. Those may have different outcomes with a treatment with a complete response or non-complete response, and that can certainly affect that.

And it's also important to point out that there are some differences in the clinical trial methodology and designs of these trials, even though BCG-unresponsive disease was trying to harmonize the inclusion criteria. There are still differences between trials, between the windows of exposure. Some trials use different treatment failure definitions.

Sometimes prostatic urethral recurrences aren't considered a treatment failure for some therapies, although that often still results in a patient needing a radical cystectomy. And some of the newer trials are typically allowing re-treatment, and how do we integrate that into our assessment of these treatments?

So with regards to the FDA-approved treatments that we have, the first one across the finish line, I believe in 2020, was pembrolizumab. And in the BCG-unresponsive CIS cohort, the 12-month complete response rate was about 19%. And of course, this is an intravenous treatment.

And the major issues with that are the toxicity—the immune-related adverse event rates of 10% to 15%, which I think is unacceptable for the vast majority of patients with non-muscle invasive bladder cancer, especially those with noninvasive disease.

And the other issue we have is there's no pre-treatment predictive biomarker such as PD-1 or PD-L1 expression to help guide who may or may not potentially benefit from these treatments. Nadofaragene was the next one to cross the finish line, and its response rate at 12 months in a similar population was about around 20% as well.

But again, cross-trial comparisons are difficult to make in this disease space. And I think, again, the issue with this treatment that doesn't really get discussed much is that there's underrepresentation of T1 with CIS. There were only five patients in the entire cohort, but out of those five patients, three of those patients with high-grade T1 with CIS progressed to muscle-invasive disease or metastatic disease. So that's 60% of those participants with high-grade T1 plus CIS.

And so with my patients, that's certainly a caution that I would have with those, and hopefully with the phase 4 additional data that will be garnered, that will be sorted out a little bit better. But I think that's part of the issue that we have in this disease space.

And there's, of course, also no pre-treatment biomarkers. There are some post-treatment biomarkers that have been published on, but in my opinion, those are largely unhelpful because we're trying to make the decision about which treatment to give to the patient up front.

And then, of course, the drug that you were heavily involved with, Sam—the BCG with the IL-15 superagonist—and the 12-month complete response rate, if you exclude the patients who were retreated, because this trial did allow re-treatment, if you exclude those retreated, the 12-month complete response rate is around 33%.

And I should highlight that these are Kaplan-Meier curves taken from all these publications, and the red lines are placed by myself to indicate what the actual disease-free recurrence rates are typically. When we report out these trials—and this is what was discussed in the session—is we report a response at any time, and then the durability of that response.

But when you're counseling patients, they want to know what will be my chances of recurrence in a year from now and two years from now. It's very hard when you're counseling patients about what's their chances of a response at any time, and then maintaining that response. So that's why I emphasized that the 12-month CR rate is about 33%. But if you allow for re-treatment, that will certainly go up, and it's typically about 45%, if I remember from the trial itself.

The issue with the IL-15 superagonist is the cohort that was the single-arm study. Only 20% of those patients had a complete response at three months, so 2 out of 10. And that arm, that single-arm cohort, closed due to futility. And I worry about that because there are very few treatments that have essentially no activity as a monotherapy but then contribute significantly in combination.

And so, of course, there is this IL-15 superagonist plus BCG. And we don't know what the relative contribution of BCG is in this particular study because it's the single-arm study. In about 50% of patients had BCG-relapsing disease. We know that many of those patients may potentially benefit from additional re-treatment from BCG. We don't know what that number necessarily is. There's some data from MD Anderson and others that suggest that in carefully selected patients, there could be a substantial proportion that do benefit.

As you know, there is an ongoing randomized trial of BCG with or without the IL-15 superagonist in BCG-naive patients. And so that will really help sort out this question. And again, there's no pre-treatment biomarkers at this point in time for any of these treatments. So that's definitely an area of need.

Also, what's currently available and many people consider the de facto standard—and we certainly use it a lot here at MSK and other academic and community sites are using it pretty heavily—is gemcitabine and docetaxel. And I always refer to this as the elephant in the room. And that's because the way that this drug combination was developed was not in a traditional phase 1 followed by a phase 2 study to demonstrate both safety and efficacy, but also whether or not it's better potentially than either gemcitabine or docetaxel alone.

Instead, it was retrospectively developed out of the University of Iowa, and the multi-center publication, which again is retrospective, looks really promising. And in many ways, it looks almost too good to be true with complete response rates in the CIS population of around 60% at 12 months. And a lot of those remain very durable. And I always emphasize that this is retrospectively developed.

Now, this is data from my group at MSK. This is just looking at our experience with single-agent gemcitabine, where patients were treated in the more historic BCG-refractory trials versus post-trial off-protocol where it was just offered to patients as standard of care, essentially.

And when we look back and we look at this BCG-unresponsive CIS, strictly defined criteria, what we find—again, small numbers—but what we find is the patients who were treated on those clinical trials that meet the BCG-unresponsive CIS definition, they have response rates or complete response rates of around 25% at 12 months.

Where those that were treated off-protocol, same criteria on unresponsive CIS, but not enrolled on a clinical trial—those patients actually do quite well. And again, small numbers, but not that dissimilar to gemcitabine and docetaxel. And so to me, this highlights that retrospective studies are not the same as prospective clinical trials.

There's some selection bias that goes into that. So I think there is some uncertainty over gemcitabine and docetaxel. And I emphasize that I worry that it's become this de facto standard, but we actually don't know how well it truly performs, and it needs to be further investigated.

In the session, we didn't overemphasize this because these drugs are not yet available, but we expect that they will be available within the next year or so, year or two. And that's the TAR-200, the pretzel system so to speak, that delivers gemcitabine continuously. I didn't go into much of the efficacy data, although the complete response rates at any time in this look very promising.

But I do worry because there has yet to be an actual head-to-head comparison with standard gemcitabine, with or without docetaxel. And when we're talking about sequencing patients, one of the issues that came up is, how does the sequence of therapy factor into things? Does cross-resistance potentially develop if one's getting treated with GemDoce and then goes on to the TAR-200 or vice versa?

And I also think there are some logistical problems as well with the TAR-200 system or anticipating logistical problems in terms of having to do cystoscopies every three weeks or so to remove and exchange these. And obviously, this is going to be fairly costly compared to standard intravesical gemcitabine.

So at the panel session, at the meeting itself, we saw a late-breaking abstract that was reported on the cretostimogene product—so the oncolytic virus. The data that I'm showing right now, and I presented, is last year's data. So the updated data will be in the late-breaking abstract session.

And the complete response rate at any time was reported at 76%, which seems very promising indeed. But my enthusiasm is a little bit tempered. When you look back at the phase 2 trial with the same drug, if you parse out what the numbers are specifically for that in the phase 2 study, which was both papillary and CIS for BCG failure, the complete response rate at any time was 65%.

And you look further at the 12 and certainly the 18-month CIS, those did not remain durable. And again, roughly around 20% of patients are experiencing durable response, and 80% of patients are having recurrence beyond 12 months or so.

So I know of no other example where the phase 2 studies look worse—the phase 3 studies look better than the phase 2. Usually there's some regression to the mean. So a lot of this makes me think that maybe we're just getting better at running single-arm non-randomized trials. And that's certainly a controversial thing to say. But it's not clear to me that these drugs are necessarily better than the currently available drugs.

And I do think that the time has come to move on beyond single-arm non-randomized trials. I think we're dealing with two sort of competing problems in the BCG failure space, and that's, we have all these newly approved FDA-approved drugs that were approved based off these single-arm non-randomized trials.

We don't know the comparative effectiveness between them. We actually don't know their true clinical utility compared to existing treatments like intravesical chemotherapies. I think this all needs to be sorted out. And there's a lot of uncertainty over what the best treatment option is for BCG failure. And when there's uncertainty, the best way to address that is through a rigorously designed randomized trial.

And I think that's also an opportune time to have some integration with biomarkers to help guide between treatment selection amongst those. But the question then is—and this is what the panel discussion also dealt with—is how do we sort out these randomized trials? How do we conduct them? There's no real incentive from pharmaceutical companies to run the trials. They already have FDA approval, their drugs already on the market. There's no incentive for them to compare it against other available drugs.

And I worry that the cooperative groups don't necessarily have the agility or nimbleness to conduct such complicated trials. They tend to move very slow. And this is a very fast-moving space, and there's numerous different options that are available that need to be treated and assessed. But at the same time, as we discussed in the session, we still need better treatments for our patients.

And so there's an endless amount of potential combinations that may be needed that we had discussed as a group. There's also some newer treatments that are more rationally designed that are coming up as well. And I think single-arm non-randomized trials should go back to what they were initially designed for and what they're used in most other oncologic areas, and that's for signal finding for drugs that look promising, that then warrant randomized controlled trial evaluation.

And then we had discussed briefly in the session as well that there may be a role for some biomarkers as more of a surrogate endpoint to get an early signal of go/no-go with the drug to move forward. And as I continue to emphasize, and emphasizing now in this presentation now, I do think that we need to start thinking more about stage-specific clinical trials as well. I think high-grade T1 with CIS, those patients are underrepresented in most trials.

And I think they need different treatment than patients with noninvasive disease. And I think high-grade T1 plus CIS is probably the ideal cohort for integration with an immune checkpoint inhibitor or with patients with noninvasive disease. To me, that makes more sense for an intravesical-only based strategy and potentially avoiding the toxicity of a systemic therapy. So that's just some of the points that were discussed in the session. And thank you for the opportunity to present that here.

Sam Chang: Gene, that was an incredible overview that was presented with—just as any good presentation raises more questions than provides answers, which is, I think, the hallmark of any good presentation. And there are so many possible questions that we could go into. But I'm just going to focus on a couple.

One is, one of your early slides looking at urology dependence, pathology dependence, disease dependence, and the variations that each of those can contribute to results of trials, to enrollment in trials, etc. And I think that point is under-emphasized and actually understated in all these discussions that we have, because your point regarding how we really don't know what we've studied, I think is really pretty much on the mark.

When you look at how we limit those variables, give me some ideas of what you think we should do as we start trials for this. Should we always have central pathology review? Should we demand blue light or only white light? Or tell me some integrations that you would want to have for these studies to try to decrease that variation.

Eugene Pietzak: Yeah, I think that's a great question as always, Sam. I think what we're in the need for is what truly is minimal residual disease in non-muscle invasive bladder cancer. I think the FDA has set the pre-treatment CIS based off a tissue biopsy as a patient having minimal residual disease, and TURBT alone cannot eradicate that.

I personally don't agree with that. And I think other folks who emphasize the role for TURBT—I've trained for. I think there's a lot of patients out there where you see multifocal CIS, and it's diffuse, and it's throughout. And you know that there's still in the bladder afterwards.

But I think there are definitely a lot of patients where it's a little bit of maybe an area adjacent to the tumor itself. You're using your enhanced cystoscopy. You're not seeing it anywhere else. You're taking potentially random biopsies of anything that looks marginally suspicious. And those are all coming back negative. And you just have that CIS just adjacent to the tumor itself. And those patients, whether they have diffuse CIS or if they just have tumor-associated CIS, they both qualify for the same trial.

And so I think what we need is we need more objective measures of minimal residual disease. And I think—and this is going to require some rigorous development, I would say—but I think probably the path forward is more in the urine itself. And I think whether that's some urinary tumor DNA-based biomarker—I think that definitely is an area I think that has the most promise.

But just like the circulating tumor DNA studies, I think we have access to the urine itself. That's after your TURBT, before you're starting the treatment. And you could better assess, does that patient currently have cancer or not? And it's very complicated, without a doubt. We know that there are somatic mutations in normal—quote-unquote "normal"—urothelium.

We know that there's mutational changes in the urothelium adjacent to the tumor. And I always teach the fellows that work with me that the closer you are to the actual visible tumor, the more pre-malignant molecular changes there are. And that's why I emphasize always being aggressive in your fulguration in the sites around there as well. And so that all impacts that to a certain extent.

So what is that threshold for residual disease or not? But I do think that that's probably going to move towards a better objective measure of what patients are starting with, and you could potentially dynamically assess them and look for reduction in the variant allele frequency, for example. And if you see an increase, then it could be occult disease, etc. So I think that to me is the most exciting area. But that first needs to be prospectively validated before we could integrate it in.

Sam Chang: I think that's so important, because as I get more gray hair, I realize—and I think personally—that there's a large majority—I'm going to say majority—of patients with CIS that will never die of their disease. That for some reason, they have the CIS. And we try these different treatments. And they don't progress. And we know they don't all progress. But I think the majority—but then there are these other actors, patients, that in fact do develop this progression of disease.

And the work that you're doing in terms of genetic evaluation both of upper tract and of bladder cancer hopefully will give us ideas and signals regarding who are those that are more at risk versus those who aren't. And so that over time we have all these different treatments. And the comparisons are going to take place. And I think it's going to come down to a variety of efficacy.

Yes, an incomplete evaluation of the efficacy based upon the studies that have been done, but also convenience, cost, etc. And so, how we sequence those, how we do immuno versus chemo—we're learning in kidney cancer that following certain interventions with IO-TKI, to try to follow again with them, are not as likely to be successful. And is that going to keep pace with bladder? Well, it doesn't seem to be, since we're studying them in BCG-unresponsive, and we're trying other immunotherapies. So a lot to learn for sure.

If there was one thing, Gene, if you had $100, and that's all we had, and you had to spend it on research in the non-muscle invasive area, where would you go? I'll make it easier for you. I'll throw my hat in there. And I haven't pre-thought this, because I just thought this would be a question for Gene.

I would spend it in exactly in the area you discuss in terms of marker, and marker prediction, and prognosis in terms of disease, disease kind of danger. And that might be in urine, that might be in the tissue itself. I think how we better risk stratify will be essential for any treatment down the line. What would you spend your $100 on?

Eugene Pietzak: Yeah, I mean, $100 is a little limiting, Sam. But it's funny, it came up in the meeting a couple of times. Whenever I'm writing grants and people are writing grants, we always talk about how bladder cancer is the most expensive cancer to treat from diagnosis to death. And we have all these newer treatments that are coming, that are available. And the price tag on them—especially compared to gemcitabine or even BCG—are astronomical.

So I think in terms of where's the biggest bang for the buck, I think we also haven't really been utilizing the tools that we had—the existing tools—as well as we possibly could have. And I think enhancing the effectiveness of BCG, for example, or better assessing, say, GemDoce. And I think, again, I agree with you. I think potentially the ways to do that may be through some surrogate biomarker or surrogate endpoint like urinary tumor DNA.

And that, to me, I think is going to be in the next few years the most exciting way to quickly get an assessment rather than these adjuvant trials where we're waiting for things to read out. I mean, in BCG-unresponsive disease, the unfortunate situation is most patients recur. But if we're able to expand that to BCG-exposed and other BCG failure states, and potentially even BCG-naive patients as well, I do think that would be our biggest bang for the buck. And trying to untangle and sort out all these new tools that we have that we don't know which one's better for which particular patient.

Sam Chang: Great points, Gene. It was great talking to you. Great session at the SUO. And we look forward to touching base with you again soon. And thanks again for spending some time with us.

Eugene Pietzak: Of course, always a pleasure, Sam. Thank you very much. Any time.