Ashish Kamat: Hello everybody and welcome to UroToday's Bladder Cancer Center of Excellence. I'm Ashish Kamat, urologic oncologist in Houston, Texas. And joining me today is Dr. Peter Black, who really needs no introduction. Dr. Black is really a true expert in the field of bladder cancer. And when we saw all this data coming out about combination with BCG plus IO at the recent ESMO, and of course the publications, I could think of, no better person, Peter than you, to join us in a discussion and chat with us and our audience about your thoughts on where we're headed as far as BCG plus X in BCG-naive patients. So with that, thank you for joining us and take it away.

Peter Black: Yeah, thank you, Ashish, for having me. It's been a while since I've had the pleasure. And just go over the trials a little bit and really highlight where the differences in trial design that explain the differences in outcomes. And I've titled this, Why Have BCG and IO Trials Read Out Differently? So just a little bit of review of the general trial designs, we have actually four trials, three have read out, the KEYNOTE-676 with pembrolizumab is not yet read out. But in general, they're very similar design, high-risk, non-muscle invasive bladder cancer, that is BCG-naive. So we're talking about first-line treatment. And patients are randomized to BCG induction and maintenance versus BCG induction maintenance and a checkpoint inhibitor. Three of the trials also have a third arm, which randomizes to the BCG induction only plus a checkpoint inhibitor with the goal of reducing the need for maintenance BCG and the toxicity that comes with that.

The trials have a primary endpoint of event-free survival, which is described in general terms as persistent CIS at six months. So patients who had CIS at baseline, it's still there at six months. Recurrent high-grade disease, so TA or T1, any muscle invasive disease, any metastatic disease and death. And so definition of an event is critical. You can see on the right actually the inclusion criteria I have, this is generic, but it's actually different from trial to trial. Any high-grade, of course, includes CIS. And not all the trials have this third criteria. So takeaway, one of the messages right from the start, this box at the bottom here, induction only plus checkpoint inhibitor, was negative in two of the trials, CREST and POTOMAC, so we're just going to leave that out for discussion. Maintenance BCG is super important. And what we want to focus on is a difference induction plus maintenance with or without checkpoint blockade. So three trials. These are the top-line results, event-free survival. And the CREST sasanlimab, POTOMAC with durvalumab, and ALBAN with atezolizumab. Sasanlimab is a sub-Q drug and we are not as familiar with it. The durvalumab, of course, and atezolizumab we know very well. And what we see is CREST and POTOMAC are positive trials with a significant difference in event-free survival. Interestingly, both have identical hazard ratios of 0.68, so there's a 32% risk of an event in follow-up.

And if you look at some of the numbers here, they're actually quite similar. So at two years it's basically 80% versus 85%, 85% with sasanlimab. At three years it's 75% versus 82%, so a difference of 5%, and then 7%. With POTOMAC, there's similarly, at two years, it's 82% versus 87%. At three years, 77% versus 82%. So similar impact of benefit with respect to events and similar hazard ratio. The ALBAN trial, however, was completely negative, not an inkling of a difference with a hazard ratio of 0.98. All three of these have been published. As you can see at the bottom, the CREST was presented at the AUA. So we've had some time to stew over the results since April. And then POTOMAC and ALBAN were both presented at ESMO just a couple of weeks ago. Importantly, they all show a very similar trend with respect to toxicity. So the grade three or greater treatment-related toxicity is below 10% with BCG alone in all three trials, but between 20% and 30% with immune checkpoint inhibition, which is quite high. Usually in systemic trials, we often see, or if we think of KEYNOTE-057 where it was pembrolizumab for BCG unresponsive disease, we see a rate of 10% to 15% grade three. And here it's much higher, so certainly a concern. So big question, I think, for discussion is, what do we do with the discrepancy in the results? And I think that CREST and POTOMAC have given remarkably similar results and their designs are really quite similar, so that ALBAN seems to be the outlier. And I think my conclusion of the efficacy is that there is a real signal there that immune checkpoint inhibition does improve high-grade recurrence-free survival, event-free survival. But the question is why does atezolizumab not do the same in ALBAN?

And so I've just highlighted some of the differences here. One, of course, is the efficacy of the drug. Does it make a difference PD-1 versus PD-L1? Well, we see that the durvalumab and atezolizumab are both PD-L1 inhibitors, so it's not just that. Atezolizumab, of course, has had other negative trials, and there's always this worry that it might not be as efficacious. I don't think we can really say that it's drug related at this point. There are other significant differences that might explain it. In the second row, you see, for example, sample size at ALBAN was a much smaller trial, so potentially lower power. Proportion of CIS is important across trials. And although ALBAN has relatively high proportion at 40%, it has a lower proportion of CIS only, meaning that this is concomitant CIS with papillary disease. And I think the heterogeneity in CIS and the biology of it probably impact the results. Critical is the duration of BCG. Standard of care, of course, is three years for high-risk disease. The trials have compromised with two years for POTOMAC and CREST because in the real world it's hard to give three years. But in ALBAN it was only one year, so maybe it was an under-treatment with respect to BCG. And the BCG induction-only arms of the first two trials show that maintenance is super important. Also important in this type of trial is re-induction. So if there's a recurrence, especially CIS, at three months, can the patient be re-treated? In the two positive trials they could, but in ALBAN they couldn't.

There's mandatory biopsy in a couple of the trials, hard to know exactly how that impacts outcomes. And then there is a key difference in definition of event-free survival. So for ALBAN, it includes low-grade recurrences and also upper tract recurrences. Since it's systemic therapy, the idea is if there's a recurrence in the upper tract, it should count. And both of these, we think of these as being relatively uncommon in patient with high-grade disease. I didn't actually find it in the publication, but there was suggestion that up to 20% of the recurrences were low-grade in this trial. And of course, that means that there are fewer of the more significant high-grade recurrences, which may under-power to detect a difference. It is remarkable that the event-free survival is actually uniform, just around 80% in all three trials. Although in ALBAN... Sorry, this is event-free survival at two years. In ALBAN, it's the same even though it includes low-grade disease. So it suggests that maybe it's a somewhat lower risk population in that study. So just summarize, ALBAN is smaller, higher proportion of CIS, but lower proportion CIS only, no re-induction, only 12 months maintenance, and EFS includes low-grade bladder tumors. So I think my take-home from this is that sasanlimab and durvalumab do reduce high-grade recurrences. Now, I said it was event-free survival, but if you look at the nitty-gritty, it's really the high-grade recurrences that are reduced. There's no difference in progression. We don't have any data on cystectomy.

There's no difference in overall survival, which we also wouldn't anticipate at this point. However, there's major toxicity. And it'd be nice to know which patients benefit the most that might justify the toxicity. But subgroup analyses really don't provide us any meaningful indication of which group might benefit most. And ultimately, I think we're going to be having the conversation with the patient, this is the benefit, this is the risk, what do you want to do? So shared decision-making.

Ashish Kamat: Thanks so much, Peter. I mean, again, you summarized that really well. I think first off, kudos to the investigators. Kudos to the investigators, kudos to the entire team of all three trials for not only thinking of doing it, but taking on something like the behemoth BCG and trying to go head-to-head with it, and unfortunately not succeeding in ALBAN and the other two showing some difference. So I think it's great that we have these trials read out. My take is similar to yours in many ways. I just think the trials were powered for BCG given in the olden era when only 16% of patients finished BCG. I do think within modern scopes we tend to see things better. We resect better, we choose patients appropriately for treatment better, and we can get them to more maintenance BCG. And now we're trying to add incremental benefit. And the way you portrayed it is really good. I think what's striking to me, and I'd love to hear your comments about this, is the low progression events. So for example, in CREST it was 3% or 3.5% of patients actually had a progression event. So they were all recurrences, fair enough.

But now we have multiple lines of therapy after something recurs to be able to treat those patients. So with that in mind, how are you considering... Let's leave aside ALBAN for a second, but we have two positive trials. How would you counsel a patient sitting in front of you, or how would you recommend our audience counsels their patients sitting in front of them about throwing the kitchen sink at them up front or coming in a little bit later after they have the recurrence or like KEYNOTE-676, the arm A, cohort A, which allows them to have recurrence after BCG and then come in with pembro? What's your thought process?

Peter Black: Yeah, well, it's going to be challenging. Just to your opening comment there, I'm very surprised that these are positive trials. Because I did think that the designs planned for a higher risk of recurrence, as you say, than we actually would expect on trial. The control arm rate of recurrence or event is approximately what I would've anticipated, but the sponsors anticipated lower, but somehow it's still positive. So I was the biggest skeptic, and I'm very happy to see that they are positive trials. And I think it's difficult. I think if you look in the metastatic setting, we're always looking at what second-line treatment did the patients get? So if you're giving some combination up front, you want to know, well, what did they get... Did they actually get the second line that you think they would've deserved to really demonstrate that it made a big difference? And here I think these are global trials and most patients will not have had access to any of the new drugs, other than maybe gemcitabine-docetaxel or single-agent chemo. So I'm thinking that in the context of the trials, this was their best chance, get a durable response to BCG and hope for the best. But it gets really complicated, especially in the US where there are multiple drugs.

Most places don't have those drugs yet. And there's an opportunity. So I think that with the results we're seeing with second-line treatments, things like N-803, TAR-200, cretostimogene grenadenorepvec, nadofaragene firadenovec, that it's probably not worth the toxicity in most patients as well as... Well, it's hard to say cost, because, of course, those second-line drugs are very costly. But ultimately most patients are doing well for a long period of time and don't need to go on to another line of treatment, so why take the enormous risk of IO therapies. So I think it's an important conversation to have with patients, but I don't see us using it widely.

Ashish Kamat: Yeah, I mean, again, I think like you said, it's a conversation and a shared decision-making. But if you look at the trade-off of toxicity, 20% to 30% added toxicity versus a 5% to 7% benefit in recurrence-free survival, not progression events, it's really going to be up to us to be able to try to figure out, even though the trials were not powered for subgroup analyses, try to figure out maybe with some post-hoc analyses, biomarkers, which patient might benefit most from the added systemic efficacy of these agents. I'm not asking you to predict or try to tell us anything. I'm not going to hold you to it, but any insights from your knowledge of biomarkers and things like that that you could say, "Well, if a patient has X, Y, Z," say high TMB, something like that, you might consider this upfront or are you still thinking of it completely as an open question?

Peter Black: I mean, I think it's an open question. I think it's always interesting to say, well, does PD-L1 immunohistochemistry make a difference? Does TMB make a difference? But it's been such a mess and everything else that we've done so far that I don't really anticipate it making a difference. If we look at urine tumor DNA, could there be a mutational signature that somehow indicates higher risk that benefits or higher risk that doesn't benefit that needs to go straight to cystectomy, something like that? I think there are a lot of questions. I think our natural tendency will be to think, if it's a patient who has higher risk disease, maybe heading towards the very high-risk disease according to the EAU classification, and maybe we'll treat those. But we don't know that those are ones that benefit the most. I think a lot of us would probably not want to treat a high-grade TA right now with the combination if you look at the data. It's not clear, but I think we know CIS certainly, diffuse CIS, can have a bad outcome, and high-grade T1 can have a bad outcome. So I think we would probably concentrate on those two. But micro biomarkers think it's... Hopefully something comes up. Hopefully the sponsors of these trials take advantage of these amazing trials to do some of the molecular work.

Ashish Kamat: And I think we're going to see at EMUC next week, there are going to be some presentations on looking at subcategory, very high-risk in these studies. So hopefully there'll be some signal there. I don't have privy to that, but I'm sure we'll at least learn, A, can we predict, or no, we can't predict. My thought, in closing, is that really progress in this field is going to depend upon two things. Number one, Merck, in the US at least, coming, and North America, giving us enough BCG that we can actually treat patients the way we need to treat them. And of course, just having combinations that we can put together doesn't mean we should use it in everybody. We really need to be able to personalize the therapy for the patients. With that, Peter, thank you always. Any closing thoughts for the listeners?

Peter Black: Yeah, I mean, it's super exciting times. And I think urologists have to be open to changes in practice and embrace it. And non-muscle invasive bladder cancer is looking quite multidisciplinary these days, and that's unusual for us, but we need to adapt.

Ashish Kamat: Absolutely. Thanks again.

Peter Black: Thank you.