Renal Cell Carcinoma — Thomas Powles, MBBS, MRCP, MD

PROF POWLES: I’ve been given the opportunity to talk about Year In Review about clear cell renal cancer, particularly around advanced disease, which I’m going to focus on.

I’m Tom Powles. I’m the Director of Barts Cancer Center.

I will talk very briefly, if I may, for a couple of minutes about the adjuvant space. I think is really important. We saw an extraordinary array of data at ESMO 2022, which was presented by a number of different investigators. We saw the previous year, and this update analysis for pembrolizumab versus placebo, which essentially showed a 37% reduction in the risk of progression, of disease-free survival, in the adjuvant setting. This is a positive, randomized Phase III. We also saw approximately a 50% reduction in the risk of death, early analysis, immature analysis, and this is with a 10% chance of life-changing toxicity: diabetes, pneumonitis, liver changes.

So this is a very positive trial and it’s got EMA and FDA approval.

This data is consistent. And if anything, the numbers are getting better, not worse, Neil. I think they’re the same, actually.

But it then came as a surprise to us, seen published in the Lancet, a negative atezolizumab study with almost identical design. Now, the conclusion may be that PD-L1 is less active than PD-1 in this setting. And I’m happy to have that discussion. Nevertheless, we know atezolizumab has response rates, in fact, it has CR rates of 11% in PD-L1-positive first-line disease. So atezo does have activity in this space.

And then we saw a negative ipi/nivo trial, which was a real surprise. And, indeed, we’ve written discussions and editorials around why that might be the case. Why would an ipi/nivo trial be negative if a pembro trial was positive? Because in advanced disease, we think ipi/nivo is at least as active as pembrolizumab plus lenvatinib. So, single-agent pembrolizumab should be inferior, I guess, to ipi/nivo.

What may be the explanations? Well, we did see more toxicity for the combination. And actually, we know patients in the adjuvant setting are less tolerant of toxicity. So that might be an important driver. The second issue is only 6 months of therapy was given in this setting rather than a year, and that’s relevant. And ipi was given 6-weekly rather than 3-weekly, as in the pivotal trials.

So different regimen. Different duration. More toxicity. And this may all have contributed. Nevertheless, we still would have expected the curves to go apart a little bit. And we didn’t see that. And so, therefore, we have to question the role of CTLA-4 in this setting. Is CTLA-4 doing anything? Is the ipi helping at all? I don’t think it is helping; otherwise, we’d have expected the curves to go apart.

And then proactively, we need to think about things like is pembrolizumab — why is that performing so well where other drugs are struggling?

And I think there are 2 really important trials we need to see in the future. The first is as a third arm of the ipi/nivo trial, which looks at single-agent nivo, and that will clear up the nivo question. And then, of course, the second really important trial, the 8Y8 study comparing ipi/nivo with nivo in the advanced-disease setting, which will tell us what the contribution of ipi is in advanced disease.

DR LOVE: So one question about the use of pembrolizumab in the adjuvant setting. What is your bar to use the drug? I think the way it’s approved was kind of based on entry criteria to the study, I think. But I guess my question is would you expect benefit for patients who are at lower risk? And do you offer it to patients? Who do you offer it to?

PROF POWLES: So I just offer it to patients who’ve completed the inclusion criteria of the study. The majority of patients in the study were intermediate-risk. There was an M1 NED population, those patients that have metastasis removed. It was only 6% of patients and the benefit was irrespective of that. And then I think the third point is that when we look in detail at those subgroups, there were benefits across all 3 subgroups.

Now that includes T2, G3 and 4. It includes T3, T4, N1, and that population we’re offering therapy to.

DR LOVE: Do you think that patients who are lower risk would benefit from the drug?

PROF POWLES: I’m not offering lower-risk patients the drug and I’m not offering the papillary cancer patients the drug. And I’m not super keen in going into this lower population.

The risk of life-changing toxicity is about 10%. And when you look at that, and then you look the risk of relapse being low, the benefit-risk ratio then becomes pretty gray. I think you’re entering into really dangerous territory in that population.

DR LOVE: So the next paper we were curious whether you have any thoughts of, is the ASCO Guideline for metastatic renal cell. I don’t know if anything you want to say about that, whether that lines up with how you approach the disease.

PROF POWLES: So I phoned Kim. We had a conversation. I’m involved in the ESMO Guidelines, which are different from that. And we are pulling together those guidelines.

We’ve gone in slightly different directions. We don’t believe that a tissue biopsy is, in these guidelines there is a feeling that we should always be biopsying new stages of treatment. We feel that relapse within 2 years of the primary tumor in areas consistent with advanced kidney cancer is important — is enough. And the reason why that’s the case is, it can sometimes be very difficult to biopsy small lung lesions. And it can be difficult to go in and do biopsies. And in fact, it can take a long time to do that and it can put patients in harm’s way.

So, we believe that biopsy is essential at baseline. And it’s important for late relapse. But not all patients need a biopsy.

We agree that cytoreductive nephrectomy may be offered in select patients with intermediate-risk disease. But as a rule, nephrectomy and the role of nephrectomy is relatively small.

We agree that surveillance after nephrectomy is potentially attractive for some patients. We also agree that the patient with favorable-risk disease, these patients, as we described before, the favorable-risk patients, VEGF-TKI/IO combination is the standard treatment. But we actually are more liberal now around the role of sunitinib in these patients.

And that, I think, is what the last point is also saying is, in some patients, monotherapy with VEGF-TKI is an option. We’re not supporting monotherapy/immune checkpoint inhibition in any patients. We’ve not seen data to support monotherapy. So for immune therapy it’s ipi/nivo. For those patients with poor performance status or other comorbidities, single-agent VEGF-TKI therapy is relevant.

We’re no longer recommending IL2. I think there was that recent randomized Phase III with pegylated IL2 versus sunitinib, which was the first negative randomized Phase III trial that we’ve had. So that, for me, turns me off single-agent VEGF-TKI and turns me off on IL2 on both of those. So we’re avoiding those therapies.

On the whole, more similarities than differences between the United States and Europe. I think we’re probably slightly more restrictive and we’re doing less interventional biopsies.

Some updates from some of the pivotal trials around advanced disease. And it’s fair to say that every year 2 or 3 of the trials are updated. We’ve got into the habit of looking at the latest data cut from the 4 practice-changing data sets. One of those 4 data sets is the cabo/nivo data set. But it would be fair to also say there’s len/pem and axi/pembrolizumab as well, as well as ipi/nivo, which are all essentially interchangeable in intermediate- or poor-risk disease.

The good-risk space is complicated, and I am going to talk a little bit about that today.

I’m going to talk about the update of CheckMate 9ER that was presented at GU ASCO in 2023, just recently.

Essentially, cabo/nivo continues to outperform sunitinib, as you’d expect. The PFS of these data sets never changes very much. This is a hazard ratio of 0.58. All of the VEGF-TKIs outperform sunitinib significantly in this environment. Lenvatinib and pembrolizumab is the most successful at that, hazard ratio about 0.45. Here you’ve got a hazard ratio of about 0.6.

But we also have this significant overall survival signal. And with this later data cut of over 40 months, getting towards 4 years, it’s 44 months, just shy of 4 years, we can see a significant benefit. These consistent benefits are seen across all of the trials. Now there’s a debate about whether or not 1 trial is better than another, 1 combination is better than another. But this cross-trial comparison is flawed because all of the control arms performed slightly differently in each of these trials and, indeed, access to subsequent therapy is studied differently in each of these trials, all of which will affect overall survival.

Actually, when you look at the detail of this, they’re all coming in a very similar points, the intermediate- and poor-risk population. And, therefore, it remains the case that these 4 combinations are largely interchangeable in intermediate- and poor-risk disease.

The one outstanding issue is around the favorable-risk population. Axitinib and pembrolizumab, when it was first released, its first data cut with only 12 months of follow-up, showed benefit across the board, favorable, intermediate and poor risk. And that resulted in not only a broad label, but broader guidance approvals, so ESMO guidance, NCCN guidance, ASCO guidance, approved in intermediate- and poor-risk patients with Level Ia evidence.

You can see from here, the overall survival benefit in the favorable-risk patients is not convincing. And, indeed, as time goes by, for this combination, but also axi/pembro and also lenvatinib and pembrolizumab, the hazard ratio in the favorable-risk population remains at about 1.

While the response rate for the combination is higher than sunitinib. And the PFS for the combination is significantly better than sunitinib. There is uncertainty about the survival signal.

And actually, were having a bit of a discussion in the academic community about whether you have to give the combinations or its acceptable to give single-agent sunitinib.

We move onto an update for the CLEAR trial. This is the most recent combination to come out of lenvatinib and pembrolizumab. Again, progression-free survival was the primary endpoint. This most recent data cut shows us data with a slightly shorter follow-up actually, than cabo/nivo and axi/pembro. Nevertheless, you can see the overall survival hazard ratio here, again in the mid-0.7s — 0.72.

Similar to what we saw previously, slightly shorter follow-up. You’ll notice the control arm performing slightly better. This group had the highest proportion of good-risk patients and know sunitinib performs best in this group. That’s probably why the control arm is performing slightly better.

I’ve not shown the progression-free survival or the response rate, but the response rate for this combination is in the high 70% PFS and less than 0.5. But the overall survival signal remains very strong for all of these.

Here is the PFS signal which, as I said before, was perhaps the best PFS signal we’ve seen for any of these combinations.

So I think that the conditional survival of ipi/nivo versus sunitinib I think comes as no surprise.

But what happens is, for the winners with ipi/nivo, they go on and do very well. And that’s true of the winners for all the VEGF-TKI — all the PD-1 combinations. The most important thing that’s happening in kidney cancer in the last 5 years is we’ve managed to essentially put a group of patients into long-term durable remission who appear not to relapse. And that’s what we see from this paper.

And that’s probably about 20- or 30% of patients going into this durable remission. And we’ll see from some of the PFS curves, that that remains to be the case. And, indeed, in the most recent analysis, the 5-year survival of these patients is about 50%.

And here you can see the survival analysis and the curves going apart and staying apart.

The one point, Neil, I’d like to make here if I may, and I’ve been making this for some time now and I don’t know how popular I am, but at 60 months, at 5 years, you can see the PFS for ipi/nivo is 30%. But that’s not 30% of the ITT population. You can see the ITT population at about 600 patients, and 30% of that 600 patients would be 200 patients, but we’ve only got about 50 patients at that 3 months. So essentially, we’ve got a lot of heavy censoring on this curve.

So for those patients who remain on trial, which is only about 10% of patients, 30% of those are progression-free. But there are many patients who are being censored out before. That doesn’t mean the progression-free survival is not 30%, it doesn’t mean it’s 10%. But it means there’s a lot of uncertainty because we’re losing a lot of patients along the way. It could be higher than 30%. But it means there’s uncertainty. And that’s really important.

This is a curve I’ve made myself, and I apologize for that, the triangles are ipilimumab and nivolumab, and the stars are VEGF-TKI/IO combinations. The green is cabo/nivo. The X-axis at the bottom is the time of follow-up and the OS hazard ratio is on the Y-axis. Each of these points is a data cut that we’ve seen — each is a publication. You can see how many publications there’ve been. No wonder it’s confusing. And the large green star in the middle is the most recent data cut, which I just showed for cabo/nivo.

And you can see here, actually what the point this makes, is that as time goes by for the VEGF-TKI/IO signal, the initial fantastic hazard ratios for survival is in the 0.4s and 0.5s, have come to land in about the same place for the intermediate- and poor-risk population as the nivo/ipi population. Now, ipi/nivo as the longest follow-up.

What does this mean? Well, it’s very simple. It means the survival signal for all of these 4 combinations is about the same at the moment. And as time goes by, they seem to track each other. And, therefore, when someone comes out and says, this is clearly the best combination, or this has the best survival signal, or this has the best landmark, I’m bought into that concept. I think we’re showing about a 30% reduction in the risk for these combinations — for all of them as time goes by over sunitinib.

The next important step in the first-line setting is the COSMIC-313 study. This study was a randomized, Phase III for cabo/ipi/nivo versus ipi/nivo. It’s an intermediate- and poor-risk patients with front-line clear cell renal cancer. The reason we picked that group of patients is that is the standard of care for ipi/nivo.

We know this is a positive trial with a hazard ratio of disease-free survival of 0.73. Toni Choueiri presented that at ESMO.

At ASCO GU, we saw a further update of this data. The update is important because the initial data was only on 550 patients. That was because it was a progression-free survival cohort. I’m now showing the overall survival cohort, which is 850 patients. It also has 5 extra months of follow-up.

Here is the trial design. As I’ve said before, front-line clear cell renal cancer in intermediate and poor, triplet versus doublet therapy, progression-free survival. Primary endpoint overall survival. Not yet hit and not yet shown.

We can see here this PITT, the 550 patient population has a hazard ratio of 0.74. With 5 extra months and 300 extra patients, that hazard ratio continues to be the same and continues to be meaningful. This underpins that particular endpoint of this trial. But as I said before, we haven’t yet seen overall survival.

One of the intriguing findings on the original data set was we showed a difference between the intermediate- and poor-risk patient. The benefit was confined to the intermediate-risk patients. It was unclear about why that might be the case. Initially, it might be because the trial is statistically not powered to show the difference. And it could have been bad luck alone. And for that reason, we looked in this larger 800-patient population rather than just the 500-patient population. And guess what, the results seem consistent. It does appear most of the benefit, even with a longer follow-up and the larger numbers, is confined to the intermediate-risk population.

Why would that be the case? Well, one obvious reason would be it’s hard to give a triplet to the poor-risk patients. However, if you look here, the median duration of exposure to treatment is 11 months in the intermediate and 10 months in the poor-risk group. So they’re getting the same exposure of therapy. And actually, if you look at the average dose of carbo and the median number of infusions of immune therapy, it’s exactly the same. So we were able to give the drugs to the poor-risk patients, but they still did less well. And, indeed, you can see the toxicity for poor- and intermediate-risk populations being approximately the same, with Grade 3-related adverse events being similar in both groups.

It’s, therefore, fair to say we’ve seen this difference. It’s not statistically powered to show a difference, but it is consistent in the bigger population. And it’s not purely because we were unable to give the triplet. I wonder if the reason why this might be the case is because biologically there’s a difference between the intermediate-risk patients and the poor-risk patients. We know the poor-risk patients are responsive to immune therapy and somewhat resistant to VEGF-targeted therapy. And that may account for the difference.

Neil, I think you’ve got a question.

DR LOVE: So yeah. Frankly, this whole area has always been confusing to me and it’s just getting a lot more confusing. So maybe you can straighten it out for me.

So let me just go back a little bit in terms of risk status and benefit of treatment. So, one thing, you have 1 algorithm that kind of relates to the risk assessment, but the other kind of algorithm you hear — I hear people talking about — is just whether the patient’s symptomatic and needs a response or not.

From a clinical point of view, when you have a patient who’s presenting for first-line therapy of metastatic renal cell — from a clinical point of view — one of the messages I get is, if the patient doesn’t need a response right away, they’re not tremendously symptomatic, generally go with ipi/nivo. So that’s kind of a clinical approach that’s fairly easy to follow. But once you start getting into the risk thing, to me, it gets a lot more complicated. So, maybe you can clarify it.

So first of all, in the original ipi/nivo study where there was no benefit in the favorable-risk patients, was that a planned subset analysis? And was that statistically valid? And do you think that really is for real, that these people don’t benefit?

PROF POWLES: So it is statistically valid because it was tested formally in a hierarchical analysis. It was a stratification factor. And so, yes, the answer is it was statistically valid at the initial data cut. I have to tell you Neil, if you look at the overall survival signal of the most recent data cut, the hazard ratio is currently at about 1, 0.93. So ipi/nivo in the long-term is not worse than sunitinib. They’re about the same. And that has created some confusion.

There is some elegant work by Brian Rini and others, looking at angiogenic signatures in clear cell renal cancer. And it showed the good-risk patients have strong angiogenic signatures. And I think it’s fair to say that the good-risk patients appear to benefit the most from VEGF-targeted therapy and the poor-risk patients probably benefit the least from VEGF-targeted therapy.

DR LOVE: So are there patients with good-risk disease that you use first-line TKI?

PROF POWLES: Yes. Yes, there are. And it’s an entirely reasonable approach. Now I know there are other people who say their only chance of long-term durable remission is with immune therapy and, therefore, you have to give both. But the evidence at the moment doesn’t support that. The Kaplan-Meier curve hazard ratios are 1.0 for all of the combinations.

And so currently, you have to have the eye of faith to feel that the immune therapy is changing the landscape in the long-term. Now it might be, Neil, the 10-year data — 10-year data, not 5-year — the 10-year data, because the median survival of this group has not yet been reached, it may be the 10-year data shows actually giving immune therapy up-front was the right thing to do. But that requires the eye of faith. And at the moment, groups like the Swiss group, and other groups, are thinking about changing their guidelines to bring sunitinib or pazopanib back into that good-risk population.

DR LOVE: Interesting. And again, going back to the original ipi/nivo study, they did not see a benefit of response rate or PFS at all in favorable risk?

PROF POWLES: No, they did not. Sunitinib was significantly better.

DR LOVE: Okay. So that’s very helpful. And then the second thing is, now you’re starting to show data with IO/TKI combination showing a lack of benefit in the good-risk patients. Again, is that statistically okay? Was it a planned subset analysis? Has that been seen with all the TKI/IO combinations. or only the 2 that you were talking about?

PROF POWLES: So this is more exploratory and it’s less statistically robust. The fact that the ipi/nivo data set a precedent and looked to this, meant that we looked for data in all the subsequent studies as well, which seemed reasonable at the time.

Had this only been seen in 1 of the 3 trials, and the other 2 trials showed benefit, we probably would have said it was just a chance event. But because it’s occurred consistently across 3 independent trials in exactly the same population, it’s likely that this is not by luck alone, and, indeed, a meta-analysis of those 3 trials who show a hazard ratio of 1.0, essentially.

And so I think it’s reasonable under those circumstances to say, although it wasn’t predefined, and, indeed, the label allows you to give the doublet in this population, if a patient comes in and says I don’t want to have the immune therapy, in the good-risk population, they’re not necessarily wrong.

DR LOVE: So yeah, just to pick up on that a little bit more. So, for example, I mean when I think about good risk, the first thing that pops into my mind is somebody who’s had a recurrence after nephrectomy several years later. I don’t know, is that a typical scenario where you see it?

PROF POWLES: Yes. Yes.

DR LOVE: So my question, if you have a patient like that, and the patient is very symptomatic — so, first of all, in general, would you be thinking about a TKI with that patient? And what about if the patient is symptomatic?

PROF POWLES: So there are 2 parts to that. I think when we look at good-risk patients in today’s environment, we should be thinking about essentially 3 questions. The first is, do they need systemic therapy? Yes or no. If the answer is no they don’t, can they get away with surveillance alone? Or can we ablate or use radiation therapy? So 2 small lung lesions, 1 cm, 5 years after nephrectomy, the likelihood is they could take another 3 years to grow to 2 cm. And if that were the case, do we need to intervene immediately? Probably not. Could we get away with surveillance? Yes, absolutely. Is radiation therapy or surgery or ablation reasonable? Yes, I think so. So I think that’s the first question. Do they need systemic therapy? If the answer is no, there are other options available to them.

And then, which good-risk patients really need systemic therapy? And that’s a question, which is, if that patient came back a month later and suddenly there were 4 and not 2 lesions, you’d probably say, well, 4 lesions, not 2. We’re growing. We’re beginning to get into problems. We’ve got 4 lesions. The likelihood is this is out of control. Then the issue is, how do we get back in control? Well, you get back in control pretty quickly with sunitinib. You’ll get in control. You’ll reverse that process.

People like Mike Atkins will say, hold on a second, this patient’s now on sunitinib for the rest of their life. The answer is they probably would be. But there isn’t evidence at the moment that immune therapy will put these patients into long-term durable remission that’s better than sunitinib as it currently stands.

Now, people like Brian Rini will come back and say, hold on a minute. If you’re going to start something, at least give them a go with a therapy that may be associated with durable remission in the long-term. And we may even be able to stop the drugs 3 or 4 years down the line and we’re not there yet. And that is an entirely reasonable argument. Also, the response rate for axi/pembro is going to be more like 60 versus 30% for sunitinib or — so the combination has a higher response rate.

So, it’s reasonable to say we’re going to going for high response, better PFS, that’s why we’re giving the VEGF combination.

But in this space at the moment, it’s entirely reasonable to give single-agent VEGF-TKI, VEGF-TKI/IO combinations, surveillance, radiofrequency ablation, or even radiation therapy.

DR LOVE: That’s really helpful. Thank you very much. Please continue.

PROF POWLES: I’m going to go into the second part of this story. We’ve talked about the first-line therapy. We now need to talk about biomarkers and the reason we need to explore this field is I don’t think we’ve been super good at biomarkers in kidney cancer. We all know in prostate cancer there is really exciting data on PARP inhibition. In bladder cancer, there’s data on PD-L1, which is mixed, of course, but we’ve done lots of studies. Also, FGFR. But in kidney cancer, we’ve really struggled. In fact, we use the IMDC Classification still, which is not really a tissue-based biomarker, more a blood biomarker.

Here, we’re showing for the first time at ASCO GU 2023. Biomarker analysis for the combination of VEGF-TKI and PD-1 therapy. The story so far, which is basically derived from the bevacizumab/atezolizumab and avelumab/axitinib data sets, that showed immune interferon gamma signatures, T-effector signatures, and CD8, are all associated with response to PD-L1 therapy. Can we repeat similar experiments with cabo/nivo and come up with responses to PD-1 therapy?

And here we can see a series of gene signatures, so this is RNA sequencing analysis, a series of gene signatures on the left-hand side, associated with response, with longer PFS to cabo/nivo.

One of the intriguing issues is third down, is an angiogenic signature. That’s not a classic immune signature. And one might say, well, why is the angiogenic signature not coming up for sunitinib? That’s an angiogenic drug. It may be that cabozantinib is a better angiogenic-driven therapy than sunitinib. Because it’s important to remember that this is not just driven by PD-1 therapy. The cabozantinib has an important role to play as well.

But you can see down here, the other red markers, those associated with classic response to the combination, are not those associated with immune signatures. So interferon gamma signatures, for example, are not popping out in the way one might expect. And, indeed, we’re seeing more hypoxia and angiogenic signatures.

This is intriguing because it’s suggesting there are distinct signatures for PD-1 versus PD-L1 therapy. And dare I say it, it may even be a marker of adaptivity or synergy for the angiogenic signatures associated with cabozantinib or versus sunitinib.

Another biomarker which we’ve explored a lot with mixed results across a board spectrum of canvases, is PD-L1. We’ve seen this sort of data before actually, with atezolizumab and bevacizumab. And what we actually see is, for the immune combination, it doesn’t seem to matter if you’re PD-L1-high or not. But, with sunitinib, with VEGF-TKI therapy, those PD-L1-positive patients are performing less well.

What does that mean? Well, PD-L1 is a broad immune biomarker. And I think essentially what it’s telling us is those PD-L1-positive tumors are the more immunogenic tumors. And as we said previously in this talk, we know those more immunogenic signatures respond better to immune therapy rather than VEGF-targeted therapy. And, therefore, it should be no surprise that those patients with the more powerful immune signatures do less well with sunitinib.

CD8 had also been shown previously to be relevant in this setting. And so we measured CD8 and we measured CD8 across broad spectrum of cancer types. We’ve looked in inflamed, cold and excluded phenotypes. And in a Cox Proportional analysis, we’ve shown that CD8, and actually CD8 topology where the CD8 cells stand, are neither prognostic nor predictive of outcomes to cabo/nivo. And again, this is surprising. Because it suggests to us that this biology may be distinct from that biology we’ve seen previously with PD-L1 inhibition.

And then, of course, cabozantinib is a MET inhibitor. And here you can see MET expression. And in both groups, the cabo/nivo and sunitinib, high MET expression is associated with poor outcome; whereas low, a better outcome.

What does that mean? Well, it means that this MET story for cabozantinib is probably less relevant than its VEGF story in terms of generating responses in clear cell kidney cancer.

We now move away from first-line therapy and we move into the era of second-line therapy. This is important because there are some more trials coming out in the near future: CONTACT-03 is a study where we’re looking at cabozantinib plus or minus atezolizumab after previous immune therapy. So this is sequencing immune therapy. But what I’m talking about today is the data to support VEGF-targeted therapy in individuals whose cancers have previously progressed on immune checkpoint inhibition.

TIVO-3 is a study of 350 patients for patients who have failed at least 2 regimens, including a VEGF-targeted regimen, but many of these patients have also progressed after an immune checkpoint inhibitor. Is comparing tivozanib and sorafenib in this setting.

Tivozanib is quite a specific VEGF-TKI therapy. Here, you can see a significant delay in progression-free survival with a hazard ratio of 0.73, a 20% improvement in progression-free survival. And, indeed, in those prior checkpoint inhibitor subgroups, you can see the hazard ratio is about the same irrespective of whether they had immune checkpoint or not.

And what this means to us is sequencing VEGF-targeted therapy remains important irrespective of whether the patients have had prior immune checkpoint inhibition.

It’s worthwhile noting, before we go on to CaboPoint, that there was no survival benefit of tivozanib versus sorafenib. And, indeed, there’s never been a survival benefit of 1 VEGF-targeted therapy versus another VEGF-targeted therapy. So from my perspective, is see them as largely interchangeable.

Neil, question?

DR LOVE: I’m curious whether or not you have tivozanib available to you? And how you find giving it? I hear people saying it’s better tolerated than the other TKIs. Do you have any experience with it?

PROF POWLES: The answer is yes, it is available. And yes, it does seem well tolerated. You know, Neil, I suspect if tivozanib had been developed first and not sunitinib, it might be the drug we use most often. But most of us I think are comfortable with sunitinib, axitinib and cabozantinib, and we ‘ve got used to using those drugs.

The issue with tivozanib is in TIVO-1, the original trial, it was a front-line trial. It used sorafenib as the control arm, which was suboptimal, and it also, because it had a forced cross-over in the sorafenib arm to get second-line tivozanib, it had a survival hazard ratio of 0.23 — sorry — 1.23. Which meant that the FDA rejected it because it said you’ve done a trial, but the survival signal doesn’t look good. And, therefore, TIVO-3 was an attempt to resurrect the drug a decade after original sunitinib came out.

And so, yes, it does look like a promising drug but its development has been a bit tortuous. And for that reason, it’s not widely used.

DR LOVE: Well, the thing that I was wondering about is I think that there’s some studies looking at tivozanib with IO, and whether or not, maybe, that kind of combination might be better tolerated than the existing? Or maybe it’s too late for something like that?

PROF POWLES: Well, there’s a randomized Phase III study called N-TIVO, which compares nivolumab and tivozanib versus cabozantinib, a standard of care, with tivozanib — I think tivozanib is the control arm, as the standard of care, to see if the combination can outperform a standard of care in this environment.

So, yes, there is ongoing randomized Phase III tivozanib trials for individuals who’ve previously — whose cancers have progressed on immune checkpoint inhibition.

So it’s not the most widely used drug, but it is going through a form of revival. And there are ongoing randomized Phase III trials.

DR LOVE: Please continue.

PROF POWLES: CaboPoint was another study presented at ASCO GU 2023. It’s a study which is actually very simple in design. It’s for those patients whose cancers have progressed after immune checkpoint inhibition, either alone or with VEGF-targeted therapy.

We know from previous trials, such METEOR, that cabozantinib is active after immune — after VEGF-TKI therapy. And cabozantinib has a survival benefit compared to everolimus in VEGF-refractory disease. The question is, does it maintain response rates in individuals whose cancers have progressed after immune checkpoint inhibition?

And you can see the design of this trial, Cohort A and Cohort B, patients who’ve progressed after immune doublets or VEGF-TKI immune therapy. And the key endpoints are response rates.

And here you can see the overall response rate of 30%. And it doesn’t seem to matter whether or not patients have had immune checkpoint inhibition previously or not.

Overall, it’s fair to say that the response rates of 30% appear not that different from what we’ve see with prospective series with pazopanib, prospective series with axitinib and prospective series with sunitinib. And, therefore, this is a drug which reinforces the importance of sequencing therapy but does not necessarily mean cabozantinib is the most active agent in this setting.

Indeed, subgroup analysis for this trial shows it working with similar response rates across broad subgroups of patients.

Perhaps a more provocative trial is a study looking at belzutifan in combination with cabozantinib. And this is an important study, the reason why it’s relevant is this doublet looks a promising doublet in this disease. And you can see here, this is a study looking at patients who’ve received prior immune checkpoint inhibition and less than or equal to 2 lines of therapy.

And we can see good responses from baseline. Active responses, high response rates, which I think is attractive. You can see the adverse event profile in line with what one would expect. Grade 3 or 4 — Grade 3 or more treatment-related events, 37%. So the combination becomes more challenging.

Belzutifan is a new agent. We’re using it more and more in kidney cancer. It’s a HIF-2 alpha transcription factor. As a single agent, it has response rates of about 25%. Perhaps the most attractive component of this drug, is it appears to combine well with therapy. And I think this is important. Because we now have randomized trials of pembro and belzutifan in the adjuvant setting. We have randomized trials of lenvatinib, pembrolizumab, belzutifan in the front-line setting. And we also have randomized trials of belzutifan in combination in immune-refractory disease second-line in the immune — in the advanced setting.

So belzutifan is being actively pursued in at least 3 big, randomized Phase III studies.

I want to shift gear if I may to the last important topic. The reason this is important is papillary kidney cancer has been, to some extent, left behind. The reason why that’s the case is we in the kidney cancer community, have just used the drugs that we’ve been testing in clear cell kidney cancer and adopted them in papillary renal cancer. But papillary kidney cancer is not a strongly VEGF-driven tumor and it’s also not a strongly immunogenic tumor; it actually has a very low tumor mutational burden. And unlike clear cell kidney cancer, doesn’t have a strong immune infiltrate.

And so it does seem strange that we’ve been using VEGF-targeted therapy and immune checkpoint inhibition, particularly VEGF-targeted therapy in this setting.

Therefore, I was surprised to see that when we use the combination of immune checkpoint inhibition and VEGF-targeted therapy in papillary renal cancer we saw higher response rates than expected. Cabo/atezo, front-line response rates of 47%. You might remember that single-agent sunitinib has response rates in the region of 5%, and even cabo is only 20%. And then cabo/nivo, again 47% response rates. Bev/atezo, even has response rates which are respectable certainly, compared to sunitinib. And lenvatinib and pembrolizumab higher response rates still.

So, overall, I think it’s very reasonable to say this papillary cohort has shown these combinations are more active than we thought initially. And the combination therapies I think now need to be considered as a standard of care in this setting, although it is reasonable to say there’s a lack of randomized data and, therefore, the level of evidence remains weak. But in this area of unmet need, when response rates of over 40% can be achieved, that seems preferential to sunitinib, for example, which has response rates of in the single digits.

Neil, there’s a question.

DR LOVE: What about ipi/nivo in those patients?

PROF POWLES: Yeah. So there’s been a bit of work with ipi/nivo. The data doesn’t look as impressive as the VEGF-TKI/IO combinations. You might come back and say, hold on a second, Tom, you’re comparing 30 patients in 1 patient and 20 patients in another study. You’re absolutely right.

There is a study coming out called SUNNIFORECAST. That is a randomized Phase III trial, which is now complete, of ipi/nivo versus sunitinib as a randomized Phase III trial in papillary renal cancer. So we’re going to learn a lot about ipi/nivo in the future. At the moment, the response rates don’t look quite as high. But I think it would be premature to say ipi/nivo is not active in this setting.

I’ve got a very quick summary slide that I was going to run past you.

The first is an observation from my perspective, we’ve reached a bit of a plateau in drug development in kidney cancer. Actually, we did this about a year ago. We’ve seen a number of drugs coming through, a number of new targets coming through, which have been disappointing, pegylated IL2, for example. Some of the arginase inhibitors. Some of the glutaminase inhibitors. Some of the newer targets have not been successful in this space. And, indeed, a second generation of immune therapies: Aug-40, LAG-3, TIGIT, other targets, haven’t reached prominence and may not reach prominence in this space.

Under those circumstances, we currently VEGF-targeted therapy, belzutifan is a new part in the armory, which inhibitor HIF2-alpha, which we hope will be approved form randomized trials relatively soon. But we also need to consider with this plateau, what the next major steps in kidney cancer are. I’m going to talk about that in a second.

I think there’s also reasonable debate around the treatment of good-risk disease and the role of VEGF-TKI and CLTA4 in combination with PD-1 in the front-line setting. There are some people who feel ipi/nivo is the best treatment. There are other people who feel you have to get in control with VEGF-TKI/IO combinations.

I hope I’ve explained to you today that actually, in terms of the long-term data, there are more similarities than differences between these 4regimens. I still believe that VEGF-TKI/IO therapy has higher response rates and is better at getting in control of disease. But I do think at the same time that COSMIC-313 data does question the role of VEGF-targeted therapy in poor-risk disease.

Again, it’s important to say that’s exploratory and underpowered analysis. And this is an area which the debate will intensify over the next 6 to 12 months.

The VEGF-TKIs are very similar. This is particularly true when we sequence them after initial therapy. I’m very open-minded about second-line therapy in axitinib, cabozantinib, tivozanib, all to me seem reasonable treatments.

I think the real skill in kidney cancer is not the choice of drugs, but how you give the drugs and the support team around that.

HIF2-alpha belzutifan is the most attractive new target. The biomarker data, it appears confusing. There was actually some very nice circulating biomarker data at ASCO GU 2023. I think we’re going to end up using circulating biomarkers in kidney cancer in the near future.

As I said before, new combinations, and I show you some data with belzutifan, do appear promising, although the COSMIC-313 trial showed us that triplets are not necessarily a homerun at this stage. We need to see overall survival before we go further.

The last thing is, in papillary renal cancer there has been a drift towards these doublets. That drift is based off single-arm trials, a lack of randomized data. I’d love to see some randomized data in this setting and SUNNIFORECAST will be the first of those randomized trials.

DR LOVE: A couple of follow-up questions. What about patients who have non-clear cell, non-papillary? How do you approach them?

PROF POWLES: So I think there was 1 subset of that, which is really important, and that’s the sarcomatoid tumors. Some patients have mainly clear cell with a component of sarcomatoid; others pure sarcomatoid. These patients must have immune checkpoint inhibition and I don’t mind if it’s ipi/nivo or VEGF-TKI/IO. These drugs are transformative for these patients. And I have a roomful of sarcomatoid patients who we’ve now cured who, 10 years ago, we gave drugs like sunitinib and they faded away within months. So transformative in that subgroup.

There are other aggressive tumors. Bernini tumors, they’re still using chemotherapy in those tumors. Other chromophobe, advanced chromophobe, advanced metastatic chromophobes, that we’re treating in a similar manner. Essentially, we haven’t done enough trials on this group of patients.

One thing I can tell you is the CheckMate 427 trial, looking at pembrolizumab in the chromophobe tumors, didn’t show a spectacular benefit. So I’m not sure immune checkpoint inhibition is the way forward in advanced chromophobe cancers.

DR LOVE: Another question is, any clinical situation where you think you’d like to use a triplet like COSMIC?

PROF POWLES: I think we need to wait for an overall survival signal, Neil. I really do. I think that there may be groups we’d like to see in the future. I think if you went back and you pulled someone off the street and said, who did we think, with hindsight, they’d be important in? I think many people would say it’s those really poor-risk patients with rapidly progressive liver metastasis get early control. But actually, in those poor-risk patients, there didn’t seem to be that benefit.

I still firmly believe that the combinations like len/pem, particularly with bone and liver disease, getting that control is really important. And it’s my preference to give those patients len/pem rather than ipi/nivo. Although, as I said before, it is reasonable to come back and say, in those poor-risk patients perhaps immune therapy is more important. I agree with that.

The final point I’d like to make in that is these are exploratory analysis. And we know from the prospective, randomized trials, that axi/pembro, len/pem, and cabo/nivo all perform well in poor-risk disease. And, therefore, it’s premature to say ipi/nivo is the right treatment for these patients.

DR LOVE: Do you know whether in any cancer, including renal cell, people have looked at ctDNA in patients who have long been on a couple of years of IO just to see what it shows?

PROF POWLES: So, yeah, the answer to that question is simple. We haven’t yet nailed the circulating biomarker question. There are some methylation signatures, which I think are more promising than the DNA signatures. And there was a nice piece of work in Nature Medicine 2 years ago regarding that. But I would say at ASCO GU ’23, we did see some really neat circulating biomarker data. And actually we saw 70% of patients with ctDNA-positive, using a personalized approach. That was twice as high as I was expecting.

Neil, this is important because there’s going to be a transformation in cancer, in key cancers. The technology for circulating biomarker is moving extremely quickly and is really promising.

DR LOVE: What do you think you’re going to see in these, I guess tumor-informed ctDNA assays, in these people who go out a couple of years on IO therapy, either those who are clinically in CR? Or sometimes you sort of see stuff and you don’t know whether it’s active disease. What do you think you’ll see with ctDNA?

PROF POWLES: So I think what we’re going to see is ctDNA clearance. And I think what I’d like to do is prospective studies where at ctDNA clearance we stop therapy. Because at the moment, it’s really unappealing — I saw a patient yesterday, he said how long am I going to be on this treatment for? We were starting VEGF-TKI/IO. And I said, well, maybe forever. At the moment, we’re not stopping these drugs. And he said to me, well, that’s not a great goal for me.

People’s aspirations 5 years ago, just give me anything because the survival was only a year when we started doing this. It’s now 5 years, which is terrific. But now I’ve got patients coming and saying, it’s not good enough just to be on treatment forever. I need to be free of treatment at some point, otherwise I’m not achieving my goals.

Urothelial Bladder Cancer — Matthew Milowsky, MD

DR MILOWSKY: So thanks for the opportunity to join you today. I’m Matt Milowsky. I’m a Professor of Medicine at the University of North Carolina at Chapel Hill and the Section Chief of Genitourinary Oncology. And I’m going to speak today. The title of my presentation is Clinical Investigator Perspectives on the Most Relevant New Datasets and Advances in Oncology. And this is the bladder cancer edition for which there has been a tremendous amount of promise and excitement.

And so I’ll start with some abstracts looking at non-muscle invasive bladder cancer. This has historically been more of a urologist’s disease, but medical oncologists are getting more involved. And so this is an abstract that was presented by Andrea Necchi at the GU Cancer Symposium, and looking at pembrolizumab for BCG-unresponsive high-risk NMIBC. And so previously, the cohort A results were presented from the KEYNOTE-057 study. And so, again, this is looking at patients with high-risk non-muscle invasive disease with CIS and they could also have papillary disease. And this was the initial presentation of cohort A. Here is just some background information about the criteria for the definition of adequate BCG and BCG-unresponsive disease. These are criteria that are endorsed by the FDA and used in the context of the trials that are looking in this clinical space. And I think probably the most important thing without getting into the details is just to point out that there’s different types of BCG-unresponsive disease. There’s those that progress, there’s the persistent high-risk non-muscle invasive disease and then the recurrent disease in spite of adequate BCG.

And so the original data from cohort A was presented and subsequently published by Arjun Balar in Lancet Oncology. And this looked at cohort A which was results from CIS plus or minus papillary tumors. And this showed a clinical complete response rate of 40% at 3 months which was deemed to be very promising. There was some disappointment in seeing that the CR rates weren’t maintained in all patients. It was about a little under 20% ultimately that remained in CR. However, certainly, a novel approach with a checkpoint inhibitor in this particular clinical disease state. And so this is the data from cohort B. So this is patients with papillary tumors only without evidence of carcinoma in situ.

And so this is the schema for the trial design. And, again, these high-risk non-muscle invasive patients. We’re talking about cohort B now, so these papillary tumors without CIS receiving pembrolizumab for up to 2 years. They undergo a series of different assessments with cystoscopies and urine cytology and imaging studies. And then, of course, if they develop recurrence or progression, they discontinue or continue treatment in the event that there’s no evidence of recurrence or progression of disease.

This is the characteristics of the patients in this study. So, again, these are the high-grade T1, Ta patients. And you can see that they’re binning to the persistent recurrent or progressive groups, so most of these patients had recurrent non-muscle invasive disease.

And so here’s the data. Here’s the Kaplan-Meier curve for disease free survival. So here, you can see that the 12-month DFS for this patient population was about 44%. You can see this falls off at 24 and 36 months at about 35% with a median disease free survival of 7.7 months.

If you look at the treatment-related adverse events, I think consistent with what we all know in terms of the use of immune checkpoint inhibitors. So you see your typical pruritis, diarrhea, asthenia, rash. You have your thyroid dysfunction. And then the less common side effects here, Grade 3/4 in about 14% of patients and about 10% discontinued therapy.

And so what does this all mean? I think it means that pembrolizumab continues to demonstrate antitumor activity, not only in that CIS cohort, but also in the T1, Ta, high-grade cohort. The safety profile is consistent with what we see for immune checkpoint inhibitors. And I think it represents a strategy for patients with high-grade T1, Ta, BCG-unresponsive disease. There’s a lot coming down the pike in this particular area. So this is hopefully going to be one of many types of approaches.

There’s been some interesting data looking at combination type approaches. This is a really nice study that was published by Noah Hahn. This actually came out of an NCI clinical trials planning meeting years ago. And this is being coordinated through or was coordinated through Hoosier Cancer Research Network. It’s called the GU16-243 ADAPT-BLADDER study. So it’s a trial of durvalumab, the anti-PD-L1 agent, with BCG or external beam radiation therapy again in that BCG-unresponsive NMIBC population. So this is, again, small numbers of patients here, but it was really looking at a multi-arm, multi-stage type trial design to look at durva in combination with other agents that are active. And so here with BCG in cohort 2. And in cohort 3, it was external beam radiation therapy. And so what we see here is that you get durva for up to 8 cycles and then you get standard sort of intravesical BCG. And in the external beam radiation therapy cohort, it was 6 Gy times 3 in cycle 1 and then evaluated with a 6-month required biopsy. 

From a toxicity standpoint, about 30% of patients across all cohorts experienced Grade 3/4, 14% with dose delays due to toxicity and 14% discontinued durva due to toxicity. And, again, I think this is once again consistent with what we see with immune checkpoint blockade. Here, if you look at the CRs, this is particularly interesting in the durva plus BCG cohort with a 12-month CR rate of 73%. Again, small numbers of patients, but certainly a suggestion that there’s activity here and CRs in nearly two-thirds of patients across the combination cohorts.

And so feasible, safe in BCG-responsive NMIBC. I think there’s encouraging preliminary activity here. But I think really most notably, this novel multi-arm, multi-stage type trial design really facilitates rapid evaluation of novel therapies, here being combined with anti-PD-L1 or, you know, other immune checkpoint inhibitors. And I think there’s justification here for using these combination approaches. And there’s a bunch of studies going on looking at these combination approaches, particularly with, as an example, BCG in combination with immune checkpoint inhibition.

Staying in the theme of non-muscle invasive bladder cancer, these are some data again from the Genitourinary Cancer Symposium. This is the THOR-2 study. This has, you can see 3 different cohorts of patients. This is the cohort 2 interim analysis. So these, again, are small numbers of patients. Here, these are patients that are molecularly selected for having FGFR alterations. And we bin them into high-risk non-muscle invasive papillary only, no CIS. Cohort 2 is much what I described earlier, the CIS with or without papillary disease. And there’s an intermediate-risk cohort which I’ll speak to. And these patients are treated with erdafitinib which is an FGFR1-4 inhibitor which is approved in patients with advanced disease. And the exploratory endpoints were CR rate at cycle 3 day 1 and cycle 6 day 1, so 3 months and 6 months. And secondary endpoints including toxicity. And so you can see the CR rate at 3 months, 100%. And at 6 months, it was about 75%. The median duration of response was approximately 3 months. And so certainly evidence that there’s activity here in non-muscle invasive disease where we know that there’s a higher prevalence of FGFR alterations as compared to muscle invasive or metastatic disease.

DR LOVE: What do we know about mechanisms of resistance to BCG? And what’s the rationale, getting back to the previous study you were talking about with durvalumab, of combining an IO with BCG in a patient who is resistant to it?

DR MILOWSKY: Yeah. So I think that, you know, it’s an excellent question. I think the mechanism of BCG is still not entirely clear. It is the first sort of immunotherapy agent in a lot of ways, and we often forget that, stimulating an antitumor response. However, the specific exact mechanism, less is really known. And I think the idea of sort of augmenting the immune response systemically while giving intravesical BCG for a local immune response is certainly the hypothesis here and the rationale behind doing those type of combination approaches with BCG even in those unresponsive patients. 

So, again, erdafitinib, activity in non-muscle invasive disease in this cohort. If you look at the toxicity profile, here you can see there are the typical erdafitinib-related adverse events. You can see treatment-related Grade greater than or equal to 3. Treatment emergent adverse events, about 30%. And if you look at the exact events, they’re what we see with erdafitinib. It’s not an easy drug to give. You have to be familiar with the side effects, the dry mouth, diarrhea, hyperphosphatemia, dysgeusia, stomatitis, the nail disorders. And so my conclusion here is that efficacy was observed with erda it looks like in small numbers of patients. The toxicity profile is consistent with the known side effect profile of erdafitinib. And I think, you know, as we look at these agents in earlier stages of disease, whether it be immune checkpoint inhibition or FGFR inhibition or others, we have to think about that risk/benefit profile in non-muscle invasive disease as compared with more advanced disease acknowledging the fact also that these patients with refractory disease will go on to lose their bladder. So it’s not inconsequential.

Cohort 3, this was an interim analysis, so a very different cohort. This is intermediate-risk non-muscle invasive bladder cancer. So these are the tumors that have a high recurrence probability, greater than 50%, but a much lower progression probability. And so these are all prior tumors. They had to be low-grade, Grade 1/2, Ta, T1, no prior carcinoma in situ. And here, you can see again erdafitinib treatment and here looking at the CR rate, about 75%. Median duration of response, about 2.8 or so months. And the toxicity profile once again is very consistent with the erdafitinib-related side effects. What I would say here is that I think as we go earlier, we definitely need to be more and more thoughtful about the sort of toxicities that we’re seeing in our patients, particularly for some of these diseases which represent more of the nuisance type of disease with continue recurrences as compared to those that progress to a more lethal phenotype.

Moving on to muscle invasive bladder cancer. This is an exciting agent that has demonstrated activity in non-muscle invasive disease. This is a really nice publication by Sia Daneshmand in Urologic Oncology that looked at this agent called TAR-200. We call it the pretzel based on the way it appears. It is a — essentially, a delivery device for whatever you want to deliver. In this case, the TAR-200 is gemcitabine. And it’s administered in two 7-day interval — 7-day cycles. In this particular trial, it looked at patients with muscle invasive disease who were either not candidates or declined cisplatin-based neoadjuvant chemotherapy and they were binned into arm 1 or arm 2 with larger or smaller tumors. Safety profile looked quite good. These are the treatment emergent AEs that were classified as TAR-200 related, so 17%. And these were procedure related. And there were really no major toxicities associated with this, really Grade 1/2 type events. Looking at preliminary antitumor activity here, pathologic responses in 40% of patients. These are less than pT2. And then here, you can see that these were about 60% in this arm with smaller tumors, that is less than 3 cm. And so I think this is certainly interesting preliminary activity for an intravesical drug delivery system and hopefully holds some promise for the potential use of intravesical therapy in the context of muscle invasive disease.

So we know that a lot of treatments in muscle invasive disease are underutilized or we can’t use them for one reason. TAR-200, novel intravesical gemcitabine delivery system seems to be well tolerated, safe in patients with MIBC. And neoadjuvant TAR-200 demonstrated what looks to be some antitumor activity in this Phase I study. So I think more to come.

And here is what is to come. This is some Trials in Progress posters that were presented at the GU Cancer Symposium. This is by Dr Psutka. This is the SunRISE-4 study looking at TAR-200 plus cetrelimab which is an anti-PD-1 agent or cetrelimab alone as neoadjuvant therapy in patients with MIBC who are ineligible or refuse neoadjuvant cisplatin-based chemotherapy. So fairly straightforward trial design with a primary endpoint of pathologic complete response at the time of radical cystectomy.

There’s also some other agents. As I mentioned, this is a delivery device system and so you can deliver different agents. In this case, it’s delivering erdafitinib. This is a Phase I first-in-humans study. So these are, again, molecularly selected patients with FGFR alterations. And they, again, are being looked at in the context of different cohorts. There’s a high-risk papillary-only cohort with prior BCG, refusing cystectomy, a papillary-only MIBC prior BCG, planned radical cystectomy in intermediate-risk and then the muscle invasive and receiving this, again, intravesical delivery system with erdafitinib, again, in molecularly selected patients. And so I think quite a bit of excitement with these novel delivery systems to be able to now send agents directly into the bladder with potential activity.

Moving on. Again, muscle invasive disease. This is some very nice data that Dr Galsky updated at the Genitourinary Cancer Symposium. This is extended follow-up results from the CheckMate 274 trial. This was a randomized, double-blind, multi-center study of adjuvant nivolumab versus placebo in high-risk MIUC. And so these are patients with high-risk disease, so they could either have received neoadjuvant cisplatin-based combination chemotherapy and had persistent muscle invasive disease ypT2 or if they didn’t receive neoadjuvant therapy, they had pT3 and/or node-positive disease. Here, randomized to nivolumab versus placebo with a primary endpoint of PFS in the all randomized ITT population and the PD-L1 greater than or equal to 1%. And so the original data was published in The New England Journal of Medicine. This had a median follow-up of about 21 months and adjuvant nivo improved the DFS versus placebo in the ITT. The hazard ratio was 0.70. And in patients with PD-L1 expression greater than or equal to 1%, it was 0.55. And so this presents the extended follow-up with a median follow-up of about 36 months.

These are some of the baseline demographic and clinical characteristics of the patients in this study. And you can see if I point out a few things, about 20% had upper tract disease. There’s renal pelvis or ureter. About 40% of patients received neoadjuvant therapy. And you can see that a fair share of patients had node-positive disease at the time of resection.

Ans so what did he show? He showed that the hazard ratios held up very, very nicely. Here with the extended follow-up, the ITT population with a hazard ratio of 0.71. In the PD-L1 greater than or equal to 1% it was 0.52. And I think this is really notable here. The nivo median DFS, 52.6 months as compared to 8.4. It’s pretty striking. If you look at the forest plot, you can see that almost all of the subgroups favor the adjuvant nivolumab. Here’s an exception with upper tract tumors. And this, you know, I think you can’t take too much into these subgroup analyses, but there is potentially some biology to explain this based on molecular subtypes and perhaps less response to immunotherapy in upper tract disease. But, again, I wouldn’t draw too much conclusion from subtype — from subgroups.

Additional endpoints, non-urothelial tract recurrence free survival, again, significant for both the ITT and the PD-L1, 0.72 and 0.53. Similarly for distant metastasis free survival. And then they looked at an exploratory endpoint of second progression free survival in this presentation. And this is time from randomization to disease progression after subsequent next line anticancer therapy or start of subsequent next line anticancer therapy or death. And here, you can see that there is again an improvement both in the ITT population and the PD-L1 greater than or equal to 1% with a hazard ratio of 0.54 here.

And so extended follow-up of the study that led to the FDA approval of adjuvant nivolumab, again, sort of establishing that these hazard ratios for DFS both in the ITT and the PD-L1 positive population, very stable. And I think that if you look at the actual median DFS, particularly in the PD-L1 greater than or equal to 1% population, again, 52.6 months, more than 6 times the median DFS in the placebo arm. That exploratory PSF end — 2 endpoint is interesting. And I think really, this just continues to establish this as a potential treatment option for patients with high-risk muscle invasive bladder cancer.

DR LOVE: So I have a question about theoretically who would benefit from this strategy. I believe the FDA indication just says it’s for patients at high risk without specifying it. But what I hear, and you can correct me if I’m wrong on that, but what I hear from investigators is they use the entry criteria for the trial. So like you said, pT2 after neoadjuvant, T3 before. First of all, is that your approach? Am I right about the FDA, incidentally?

DR MILOWSKY: Yeah, you’re correct. I think you’re right in the interpretation. I think that the — historically, as I’m sure you’re aware, what has been used in the adjuvant studies of chemotherapy has typically been that T3 and/or node-positive disease. So it’s pretty consistent. We know patients, for example, with node-positive disease post-neoadjuvant chemotherapy, if you look at 5-year overall survival outcomes, they’re close to 0. So there are certainly higher-risk patients in this group. And so I do think it’s a conversation with patients, particularly since the IMvigor010 study was a negative study of adjuvant atezolizumab. And we don’t have the pembrolizumab AMBASSADOR study yet. Dr Powles has done some really exciting work with circulating tumor DNA. And there’s a trial called the IMvigor011 study which is going to try to answer that question of looking at circulating tumor DNA to sort of select patients for adjuvant immunotherapy.

DR LOVE: So my question is theoretically, would you expect patients who are at lower-risk to also benefit from adjuvant nivo? And do you think that in some of those — do you use the trial criteria to decide whether or not they’re going to be treated?

DR MILOWSKY: I do use the trial criteria to decide whether or not they’re going to be treated.

DR LOVE: So if you have patients who are at lower-risk, do you think there are some of them who actually if you, whether or not you think it’s legitimate, to apply the hazard rate going down the way they do like in breast cancer, for example. But do you think that patients who are lower than that bar would end up with a positive risk/benefit ratio from, putting aside the issue of the negative trial, just looking at this study?

DR MILOWSKY: Yeah. So I think there are going to be some patients for sure who are lower-risk who may benefit from the use of adjuvant therapy. I think the issue is that you then get into the risk/benefit of giving someone immune checkpoint inhibition for a year with the potential toxicity. These are long conversations to have, as you’re aware, when you’re talking about adjuvant therapy with patients. And this one is particularly difficult just based on the data we have. If you look at the circulating free DNA story, it looks as though you may be able to pull out those lower-risk patients and not subject them to a therapy that they don’t need with the associated toxicity. So I’m really hoping that we have that biomarker in the future to be able to do better in terms of selecting patients.

I don’t think you rely on subgroups at this point. I think you have to be really careful in doing that. So I have this conversation with my patients with high-risk disease, but we talk about, you know, the adjuvant chemotherapy data, the other studies and the potential toxicity associated with treatment and the possibility of overtreatment.

DR LOVE: All right. Please continue.

DR MILOWSKY: So moving on to metastatic disease. There was some exciting work, updates of the IMvigor130 study. There were 2 studies that were exciting to a lot of people in bladder cancer in the metastatic space where we were looking, much like has been done in lung cancer and other cancers, at the combination of chemotherapy and immunotherapy. And one was the KEYNOTE-361 study with pembrolizumab. And this is the IMvigor130 study with atezolizumab. And so here, 3 arms, atezo and platinum-based chemotherapy, atezolizumab monotherapy arm and then a chemotherapy/placebo arm. And you can see the hierarchical statistical design of this study looking first at the progression free survival, looking at arm A compared with arm C. And then from there looking at overall survival in A versus C, B versus C and then B versus C in the high PD-L1 population. And so, again, this is looking specifically at this atezo monotherapy group. In the original analysis, there was a significant progression free survival benefit associated with atezolizumab and chemotherapy as compared to the placebo/chemotherapy although the median progression free survival was at 8.3 versus 6.2 months. So I think there was a lot of question about how clinically significant that really was, clinically meaningful. And really now, without a survival benefit associated there and the negative data from the KEYNOTE-361 trial with pembrolizumab. So that sort of drained a lot of the enthusiasm along with the JAVELIN 100 Bladder study that really established maintenance immunotherapy after chemotherapy.

But this was really interesting to see. Here, you can see the final overall survival analysis in the ITT population. And in fact, no benefit with regard to overall survival with atezolizumab versus chemotherapy.

But here it is in the PD-L1 high population. So this is the IC2/3 population. And you can see here, hazard ratio of 0.56 whereas there’s no benefit in these patients with low PD-L1. And if you remember back to about 2018, this is — sort of resembles the revision of the FDA approval of both atezolizumab and pembrolizumab in bladder cancer where they looked at some preliminary data and the indication was changed to look — to only cisplatin ineligible patients who had high PD-L1 expression. Subsequently, atezolizumab was withdrawn. And subsequently, pembrolizumab changed the indication to patients who were unfit for any platinum-based chemotherapy. But I think this does lend, you know, support for the potential benefit that can be seen with up-front immune checkpoint blockade in patients who are high PD-L1 who are not cisplatin eligible.

And so, again, I agree with the author’s conclusions that, you know, no survival benefit, but the exploratory analysis is, I think, interesting and consistent with, you know, I think what we’ve seen in the clinic when we care for some of these patients, and represents a consideration for first-line treatment for cisplatin ineligible patients although, again, we have level 1 evidence now with the JAVELIN Bladder 100. And so I think that that maintenance immunotherapy in patients who are chemotherapy candidates continues to remain the standard of care at present.

Moving on to some, again, really provocative data with enfortumab vedotin in combination with pembrolizumab. So enfortumab vedotin is an antibody drug conjugate. This is targeting nectin-4 using the payload monomethyl auristatin E. And here is data from the EV-103 cohort A. So this was published by Chris Hoimes in the Journal of Clinical Oncology. This was an expansion cohort looking at EV given on days 1 and 8 with pembrolizumab. And here, you can see the response rate was 73% with a CR rate of 15.6%. You can see the durability of some of these responses that were seen.

And if you looked at the Kaplan-Meier curves, again, this median duration of response, over 25 months. And here’s the progression free survival at 12 months and the median overall survival at 26 months. And so I think that in this dose escalation expansion cohort within A, this really set the stage for more study of this combination of immunotherapy and EV. Treatment-related side effects were consistent with the individual agents. A little bit more, we have skin toxicity in combining pembro and enfortumab vedotin, but the neuropathy, fatigue, alopecia, skin side effects. And these are things, again, that we’re quite familiar and comfortable with managing in these patients.

This was part of the multi-cohort study, the EV-103 study which is, again, a multi-cohort study that’s looking at patients in different clinical disease states with combination approaches including immunotherapy and chemotherapy. I showed you the cohort A data. This is the cohort K data. And so, again, that’s that dose escalation expansion cohort A that we just saw. Cohort K looked to establish the activity of EV/P further in patients who were in the first-line cisplatin ineligible setting. There was also an EV monotherapy group here, and patients were randomized. It’s important to note that there’s no formal statistical comparison between these 2 treatment arms, but they needed to get some additional information related to EV monotherapy in this particular patient population. And this represented a really nice opportunity to be able to do that. The primary endpoint being overall response and a bunch of important secondary endpoints as well as exploratory endpoints which I’ll come to in terms of patient reported outcomes.

And so this was the data that Dr Rosenberg presented at ESMO in 2022. Response rate, 64.5% for EV+P. Median time to response of 2 months. 97% of patients had tumor reduction. And 86% of the responses were observed at the first assessment. So I think this is really not, you know, we haven’t seen this before in patients in the first-line cisplatin ineligible space. If you looked at the EV monotherapy arm, response rate of 45.2%. And this is very consistent with the prior results that we’ve seen with EV monotherapy in a second-line+ setting in patients with locally advanced or metastatic disease.

Median duration of response for EV+P was not reached. 65% of responders were still responding at 12 months. And here are the PFS and OS curves which I think are, again, really promising. A median PFS not reached here in the EV/P arm. And if you look at median overall survival, about 22 months. And so really establishing after the cohort A data that there is this exciting activity associated. And there’s a randomized trial that’s going on called EV-302 which is looking at this in that clinical disease state as compared to chemotherapy in patients who are both ineligible and eligible for cisplatin-based chemotherapy, trying to get a handle on whether or not this could actually become a new standard of care in patients with bladder cancer metastatic in the first-line setting.

Dr O’Donnell looked at a subgroup analysis of response rate as a poster presentation at the GU Cancer Symposium. And you can see here, again, the numbers are small. These are underpowered. You can’t draw definitive conclusions here. But I think what you can say just if you sort of look here that the overall responses are very consistent across these different subgroups of patients, both for EV plus pembrolizumab where the response rates are higher as well as among the EV monotherapy patients. And these are quite consistent with what we’ve seen previously. And so, again, just establishing that there appears to be benefit across subgroups. EV/P is being further evaluated again in the EV-302 study. It’s also being looked at in patients in the MIBC setting. And we really look forward to more data to understand how this potentially could become a new standard of care.

So I presented some work looking at patient reported outcomes for the EV K group of patients. Here, we looked at patient reported outcomes using the EORTC QLQ-C30 and the Brief Pain Inventory Short Form. These were administered at baseline, weekly for cycles 1 through 3 and then every cycle until the end of treatment. And you can see the EORTC-QLQ-C30 is grouped into cancer-related symptoms, function and quality-of-life. And the BPI-SF looks at pain with different ranges. Here, higher scores representing greater symptom burden, higher functioning and higher quality-of-life. And here, higher scores representing more pain.

And here is the data from the symptom scales. And EV/P was associated with preservation or improvement in symptom scale scores. And I think most notably here, you see this meaningful improvement in pain at week 24. You do see some transient worsening of symptoms, most notably diarrhea that then returns to baseline. And it’s important to note here that the tool that we use doesn’t actually capture the cause of pain. So we know that EV causes neuropathy, but it wasn’t possible to determine specifically neuropathic pain based on the use of this tool. But it’s certainly promising to see the improvement.

If you look, again, using the BPI-SF, this is looking at pain, more than one-third of patients had moderate to severe pain at baseline. And then if you look at the improvement in worse pain, you can see that this was demonstrated quite nicely with the EV/P treatment arm. Here, you can see this meaningful improvement by week 24. And worse pain, average pain and pain interference and severity all showed improvement from week 4 through week 24.

And if you look here in the 45% of patients who had moderate to severe pain at baseline, 85.7% had a sustained improvement in pain with a median time to improvement of 1.1 months. And I do think this mirrors what certainly I’ve experienced and speaking with other investigators in the clinic where EV does help in this way with these patients who are experiencing pain with rapid responses when those patients are in fact responding to treatment. And so I think this just lends additional support to the exciting efficacy data that we’re seeing in EV-103 cohort K and A with regard to patient reported outcomes.

So moving on to, again, metastatic disease, but these are in patients who are previously treated with platinum-based chemotherapy and immune checkpoint blockade. This was the EV-301 study. Dr Rosenberg presented the 24-month findings from this study. And here, enfortumab was the randomization versus chemotherapy with either docetaxel, paclitaxel or vinflunine with a primary endpoint of overall survival. And this led to the full approval of enfortumab vedotin. And this was the 24-month follow-up data.

And I think what you can see, this was the NEJM publication with a median follow-up of 11.1 months. Here are the medians. Here are the hazard ratios, 0.7. 0.62 for PFS. And, again, data holds up beautifully. If you look at the median follow-up of about 24 months, almost identical with regard to these hazard ratios. Again, I think just establishing this is a standard of care in patients who progress after up-front platinum-based chemotherapy and immune checkpoint blockade.

DR LOVE: We actually had a case that a medical oncologist presented a couple weeks ago of a patient who had a lot of comorbidities, thought to be platinum ineligible. And the doc actually was able to get EV plus pembro and treated him. And one of the questions he had, and I’m curious in general, how EV and EV/pembro is tolerated in patients with a lot of comorbidities and very elderly. Do you have to do dose reduction? What do we know about tolerability of the combination or just EV in general in older, frail patients?

DR MILOWSKY: Yeah. I think like anything, we take the data from the clinical trials in those selected populations of patients who are ECOG 0/1 and then we go into our clinics, right? And we take care of patients who are not those patients on clinical trials. And so I think that much like any other anticancer therapy that has toxicity, we use the patient in front of us and look at their coexisting medical problems and make decisions. And so what I would say is most importantly to examine those specific toxicities that are related to that agent. And in particular, the hyperglycemia, the neurotoxicity and the skin toxicity. I think in older patients, all of these treatments can be more difficult. I think dose modification, we absolutely do. You’ve got to be on top of it. And so we either do dose modification and I think sometimes we also do schedule modification as well which is actually not really embedded in the current studies. And so we may have to omit a dose and do a 1A schedule as an example in patients just getting EV and omit day 15. We’ve even done other iterations as well based on patient tolerability. I don’t know if that answers your question.

DR LOVE: Yeah, it does. It does. And also, I’m kind of hearing — is this before the FDA right now in platinum ineligible patients?

DR MILOWSKY: It is. Yeah. So we’re waiting to understand what this is going to look like for our patients. Don’t have timelines, but whether or not this will receive an accelerated approval based on the data from cohort K is, I think, to be determined.

DR LOVE: Please continue.

DR MILOWSKY: Sure. And so moving on to targeted agents. This was our first targeted agent. We talked about erdafitinib in the non-muscle invasive disease space. Again, it’s an oral pan-FGFR (1-4) inhibitor. This was data from the BLC2001 study that was initially presented and published by Dr Siefker-Radtke. Patients with metastatic UC, had FGFR mutations or fusions. Prevalence in metastatic and muscle invasive disease is somewhere in the 15 to 20% range, significantly higher in patients with NMIBC. And so here, the response rates to erdafitinib, 40%. This was the original data with a median overall survival of about 14 months.

This is the updated data, the long-term follow here. PFS, 5.5 months. And the median overall survival was 11.3 months. Activity in patients who recurred after chemotherapy as well as after prior immunotherapy as well. And so clearly, an active agent.

It has its toxicities, as we talked about earlier. I always say to patients, just because it’s an oral agent and we’re not giving it to you intravenously, this is still something that in many ways we need to even monitor more closely. Some of the toxicities we’re less familiar with as medical oncologists like central serous retinopathy and retinal pigment epithelial detachment. And these patients need to be seen by ophthalmologists along the way for regular interval screenings. And the nail toxicities can be quite significant for these patients that often require dose holds and modification. And so, again, I think like anything, if you know the toxicities of the drug and know how to follow these patients and know how to involve the other subspecialists, you’ll be in good shape. But you definitely need to stay on top of it.

Moving on to another antibody drug conjugate, this is the TROPHY-U-01 study of sacituzumab govitecan in platinum ineligible patients with metastatic urothelial cancer progressing after prior checkpoint inhibition. So sacituzumab govitecan is an anti-TROP2 agent using a hydrolysable linker to SN-38 which is the active payload which is the active metabolite of irinotecan, a topoisomerase I inhibitor. And so the original data that we saw was from cohort A which led to the accelerated approval. This is also approved in breast cancer, as you’re aware. And here, this was data that Dr Petrylak presented at the GU Cancer Symposium looking in patients who progressed after immune checkpoint inhibition, but were platinum ineligible at the start of the study. Primary endpoint was overall response.

What do these patients look like? About 30% of patients had evidence of liver metastases. 50% received prior neoadjuvant chemotherapy. 18% received prior enfortumab vedotin. And these were the immune checkpoint inhibitors that patients received.

And so in looking here, response rate a little bit higher than what we saw in the patients that previously received chemotherapy. Here, it was 32%. There weren’t any complete responses. But there was an overall response of about 54% in patients without prior platinum-based chemotherapy or enfortumab vedotin. That is, they could have, remember, received platinum-based chemotherapy in the neoadjuvant setting here, but not for metastatic disease. Median duration of response, 5.6 months. And you can see that some of these responses are durable. It's not a lot of patients here, but I think, again, just confirming that this agent certainly has activity.

And this is the progression free survival and overall survival curves. So a median PFS of 5.6 months and a median overall survival of 13.5 months. And the follow-up here, the median is approximately 9 months. So what did I take from this? I take that sacituzumab govitecan has activity in patients with metastatic urothelial cancer.

That being said, it also has toxicity. And here, this is a different type of toxicity based on the payload. So this is what we know comes with SN-38. It’s the myelosuppression, it’s the diarrhea, it’s the nausea. And I think what’s notable here is that G-CSF was received by 18% of patients for primary prophylaxis and 26% for secondary. So over 40% of patients needed to receive growth factor support with this agent. And so that’s, I think, one of the most important take-homes with this particular agent. When you asked me about older patients who are frail, I think we really need to think about, for example, the use of this agent with regard to the myelosuppression and the potential use of up-front growth factor support.

This is the study that has not reported out yet. This is similar to the EV-301 study. This is looking at patients with metastatic disease. A large, randomized Phase III study in patients progressing after platinum-based chemotherapy and immune checkpoint inhibition, randomized to sacituzumab govitecan versus physician choice chemotherapy with a primary endpoint of overall survival. So we’ll see where this ultimately lands up after the TROPHY study.

DR LOVE: I know that typically enfortumab is given before sacituzumab. I know the data are stronger. I’m just kind of curious though in terms of relative tolerability, not just from published literature, but from your own experience from a quality-of-life point of view. How would you compare sacituzumab to EV?

DR MILOWSKY: Yeah. So it’s a great question. And I think we’re absent of a little bit of data. There was some preliminary data related to the response rates in patients who previously received enfortumab vedotin, and what that looks like maybe in the as high as 30% range. I think that my experience is that sacituzumab is perhaps a little bit more difficult to tolerate. But, again, it’s a little bit unfair because I’ve used it in patients that had been previously treated with enfortumab vedotin. But I do think that the spectrum of side effects associate with it, particularly with the GI and myelosuppression, may be more difficult in some of those older patients that you described earlier. There may be select patients for which enfortumab is just not appropriate with significant neuropathy or very poorly controlled diabetes and things like that. But I think in general the way that we do this now is certainly with enfortumab first. And, again, the strength of the evidence is more there, as you suggested.

DR LOVE: And is sacituzumab plus IO being looked at?

DR MILOWSKY: Yeah. And I’m going to get to that. I’ve got some slides. It has actually been looked at. And there’s some interesting data with regard to comparing, although not directly, the different combinations of immunotherapy and immune check — and antibody drug conjugates.

DR LOVE: Okay. Please continue.

DR MILOWSKY: Yeah, absolutely. So what about other targeted therapies? I feel like we went from real excitement with the Cancer Genome Atlas in bladder cancer where we disclosed all of these potential targetable alterations in urothelial cancer to then kind of dropping that quickly because then came immune checkpoint blockade. And we went a bunch of years with the approvals of immune checkpoint inhibitors and I kind of think we’re back to where we started. And this is the BAYOU study of durvalumab, an anti-PD-L1 agent plus olaparib, a PARP inhibitor in previously untreated platinum ineligible patients with metastatic disease. And so GU Cancer Symposium, as I suspect you’re aware, was really interesting this year, particularly in prostate cancer with all the data with PARP inhibition. Well I think there’s a potential role in urothelial cancer as well. Homologous recombination repair gene mutations are actually quite common in urothelial cancer. If you look at germline alterations, you know, germline alterations in urothelial cancer are on the order of somewhere about 14% in the Memorial experience. And if you look at what those are, those DDRs, they’re maybe on the order of 6+% of patients in terms of the BRCA1/2 and the mismatch repair defects. And so they’re there. And if you look at somatic alterations, it’s probably upwards of about 25% of patients will have some DDR alteration. This is a randomized multi-center double-blind Phase II trial looking at durva with placebo versus durva and olaparib with a primary endpoint of progression free survival. And you can see here, they looked also at survival outcomes in the HRR mutated group.

And so here’s the data. Negative in the ITT population. But this is the HRR mutated group. And here, what you can see is the suggestion that in fact if you select out patients who have homologous recombination repair alterations that you can select for potential sensitivity to PARP inhibition here in combination with durvalumab. And so I think there’s more to come here. There’s been some other data in the maintenance setting which suggests similar. And so I’m excited again to think about the use of targeted agents in urothelial cancer and potential new combinations even with immunotherapy here.

So other antibody drug conjugates, new agents. This is an RC-48 ADC. So this is a HER2 targeting agent, another ADC. This is — includes the payload monomethyl auristatin E which we’ve already heard about with enfortumab vedotin. And there were 3 abstracts actually at the ASCO annual meeting.

And the first abstract looked at patients who were HER2-high. I think getting to one of the questions or part of the questions that you were starting to ask earlier. One of the things here is, are there real differences between some of these antidrug conjugates based on the payload, and even potentially based on the linker that might pair better with different therapies? Here is just looking at HER2-directed therapy. So this is T-DM1, deruxtecan and this is disitamab which is the RC-48. And here, you can see antimicrotubule, topoisomerase I, MMAE. And maybe there are better effects when you pair these with, for example, immunotherapy. And so this was a nice slide by Dr Galsky just pointing out the antigen, antibody, payload and linker issues when it comes to these drugs.

This is looking at HER2 2-3+ metastatic UC. Response rate, 50.5%. When you looked at the high IHC 2+, FISH+ or IHC 3, 62.2%. A little bit lower in IHC 2+, FISH-negative. And here, I think very exciting activity in patients with visceral metastatic disease at 52% and metastatic disease to the liver similarly as well as patients that received, you know, prior immune checkpoint inhibition and chemotherapy.

There also does appear to be activity in HER2-low. It’s not nearly as pronounced. But here in the IHC 1+ group, you can see that there is in fact responses that are seen. And I think this is something that’s sort of panned out in breast cancer as well where you can see responses in patients who have lower levels of HER2 expression.

And then this has also been looked at in a third abstract in combination with the anti-PD-1 agent, toripalimab. And this is very striking, right? Very similar to what we saw with the combination of EV and pembrolizumab with a close to 72% response rate, again, with durability to many of these responses.

And so getting back to your excellent question which is, is there something about these combinations? Are there differences, for example, in sacituzumab and immunotherapy as compared to enfortumab or disitamab? And so here’s another nice table that looks at the response rates with these different combinations with immunotherapy. And I think what’s quite striking here is that there’s very similar response rates here. The payload is the MMAE which is this anti-tubulin agent which is quite different than what was seen with trastuzumab deruxtecan and nivolumab or sacituzumab govitecan and pembrolizumab. And so I think that this is early, clearly not direct comparisons, but maybe a suggestion that choosing the right payload to go along with immunotherapy may make a difference with regard to the sort of effect that we see.