Managing patients with FGFR-altered urothelial bladder cancer (UBC)

DR TAGAWA: Thank you very much for the invitation to give a discussion on the management of advanced urothelial carcinoma. My name is Scott Tagawa. I’m a professor of medicine in Urology at Weill Cornell Medicine, New York-Presbyterian, Meyer Cancer Center.

The update in terms of systemic therapy for advanced urothelial carcinoma is going to focus on FGFR-targeting as well as antibody drug conjugates.

We’ll start off with FGFR-targeting. I think that most of you that are watching or listening, see a multitude of different cancers. FGFR is an important pathway in a number of different cancers that are out there. In my mind 2 stand out in terms of being affected a lot: cholangiocarcinoma is one of them, not a tumor that I treat. And there’s, in part, that there’s a lot that’s out there because there is a drug approval that’s there.

In urothelial carcinoma, as you can see on the right, the vast majority of patients that present with non-muscle invasive, these papillary non-invasive tumors, the majority have activating FGFR3 alterations. But then when we look at those that we in the medical oncology world will most often see, metastatics, it’s about 1 in 5. So about 20% will have an activating alteration, most commonly an FGFR3 mutation as it comes to urothelial carcinoma, different than, let’s say, cholangiocarcinoma, which is most often fusions.

I will take a moment to say that maybe you’ll say oh, it’s only 1 in 5. It is 0 in 5 if we don’t check. So it’s one reason that we should assess somatic or tumor alterations. And I most commonly do this with panel type of testing.

So there are a number of FGFR inhibitors that are out there. Some of these are approved; some of these are not approved. This is not an exhaustive list. But what I just want to demonstrate on the slide is you can see that all of these are multikinase inhibitors in terms of hitting FGFR1, 2 and 3. There’s homology between these. so it’s very difficult when developing a small molecule to hit only 1 of them, although the next-generation is becoming more specific.

So, I hope most of you know that we actually have an approved drug for urothelial carcinoma, and I’m going to go over the data that led to that approval.

What’s called the BLC2001 trial, enrolled patients, and you can see on the left with locally-advanced unresectable or metastatic urothelial carcinoma. There was screening looking for activating alterations, again mostly mutations, but some were fusions. And then those that met the criteria got into an initial randomized dose finding, so it was actually 3 different regimens that are there, and rather than having the Regimen 3 twice, you can see Regimen 3 was picked and then that was expanded in terms of the true Phase II. On the right side, you can see the primary endpoint was objective response rate.

The entry criteria, not all of them, but you can see listed below, clinically was having basically being platinum-refractory, or if they were cisplatin-unfit, they were also allowed in. and there was no requirement or exclusion criteria for immunotherapy. This started in the kind of pre-PD-1/PD-L1 era, and then continued on during.

These are those that were enrolled. There were actually 101 that were treated, but these were the 99 that met the cutoff for the presentation at ASCO, the New England Journal publication, and were submitted to the FDA. You can see here that about 9 in 10 had prior chemotherapy. About 1 in 10 did not. And about 2 in 10 had prior immune checkpoint inhibitors. The majority, about 80%, had visceral metastasis, which is a poor prognostic factor for those who don’t treat a lot of urothelial carcinoma.

This is not the entire list of adverse event. These are those that have been identified as the importance or special interest. I would say that body surface areas, kind of as a broad aspect, so skin event. So some dry skin that usually is not so big a deal, but there are events of hand-foot. Not so different than you might see with other TKIs, but something that is there and you manage in a similar way in terms of dose adjustments as well as topical.

There can be specific nail issues, and these can be painful at times. Again, managed with dose holds or adjustments, but also, some topical care as well as antibiotics. In the trial, cephalosporins were recommended.

Something that doesn’t happen in the majority, but I know is worrisome to many practicing clinicians that are not used to it, is eye toxicity. We have more drugs out there, some antibody conjugates have eye toxicity, there’s a specific one called central serous retinopathy, or CSRR. So, because of that — and this is expected as part of the pathways — because of that, at baseline what is recommended is a comprehensive eye exam which can be done by either an ophthalmologist or an optometrist, as long as they have the equipment — that’s at baseline, monthly times 3. And then every 3 months thereafter.

I often employ an Amsler grid, which is essentially just graph paper, at specific intervals, 1 eye and then the other eye. And if any area is blurry, that’s a reasonable screening to catch this early and then send them for an extra visit. Or if they can’t get into their routine visits, then we can do that also.

Hyperphosphatemia, I’m going to talk about a little bit later, but that’s an on-pathway effect. It’s listed here as an AE, but it’s actually built into the regimen that I skipped over a little bit. We start at 8 mg. If phosphorous doesn’t go up, and there’s no other AEs, then we recommend titration up to 9. And I’ll show you some slides on why that is.

Now this is jumping to a subsequent publication. This is the longer-term follow-up where the “n” equals 1 and 1. The good thing about this, I would say the numbers are, overall, the same. So there was nothing expected. The patient didn’t fall off a cliff in terms of overall survival. And there were no new, unexpected AEs. But in terms of efficacy, and what led to the FDA approval, was this overall response rate of 40%, and that included response in the liver, response in the lung. All the different subgroups looked like benefit versus historical controls of, for instance, taxane/chemotherapy where we might expect 10- to 15%. So this led to the accelerated approval of the drug while the confirmatory Phase III is enrolling.

I mentioned serum phosphate. So inhibition of the FGFR pathway leads to blockage of phosphorus reuptake in the kidney. Now this is something that over time, you can see in that second row is transient, so, over time, there’s re-compensation by the kidney and then it tends to adjust itself, even if we don’t adjust doses. But this is what I would call an on-target effect or pharmacodynamic marker.

So in the GU oncology world, there’s precedent for giving a TKI, assessing the pharmacodynamic marker, and then adjust. So axitinib, we start off with 5 BID. If blood pressure does not go up and there’s no other adverse events, we’re supposed to go up to 7, or we go down. And that’s very similar to what was done in this particular trial based on Phase I as well as PK modeling, 5.5 was considered a prospective cutoff. And then we looked at that cutoff retrospectively.

So what you can see on the left was about a third of patients went from normal phosphorus to above — greater than or equal to 5.5 at that 2-week mark. Those that are below the line, were supposed to either up-titrate or not, depending if they had other adverse events, which you can see in the lighter green. And then at day 28, you can see that we went from about one third, close to two thirds, of those that either achieved that pharmacodynamic marker of an elevated phosphorus and/or up-titrated. What you can see in the right, to caution you as hypothesis-generating, this is prospective data. We had the prospective cutoff of 5.5, but a retrospective look at that data where you can see that those that never achieved a phosphorus of 5.5 and were not up-titrated, you can see the overall response rate versus those that achieved a 5.5 or greater and those that are up-titrated. So a higher overall response rate.

And again, hypothesis-generating, but that looked to be true for progression-free survival as well as overall survival.

So, on a practical basis, we prescribe 2 4-mg tablets. And then to go up to 9, it’s a new prescription. But we do recommend looking — I generally look at the week 2, assess how they’re doing, and then prepare for their prescription the next month with the 9-mg of maintain that 8, or if they’re having other AEs, then maybe they need to go down. But it’s something that is not just — there’s some data to back that up that it looks like there’s actually better efficacy with the drug if we follow the recommended dosing.

Not surprisingly, when we are hitting the target harder, there are more adverse events, at least a trend for more adverse events that when the phosphorus marker or cutoff was hit.

Antibody-drug conjugates as treatment for patients with metastatic UBC

DR TAGAWA: So, switching gears from TKIs targeting FGFR, to the 2 currently approved and available antibody drug conjugates, this antibody drug conjugate called enfortumab vedotin, or will maybe abbreviate as EV is what I’m used to saying, is one that targets nectin 4. It happens to be overexpressed on the vast majority of urothelial cancers. You can see on that TMA, it was 83%. But in initial development of this drug, they tested all the samples and they just kept going and decided not to restrict entry based on nectin 4 because it was the vast majority. This is the typical ADC with a protease-cleavable linker and a fairly strong payload. This one happens to be MMAE, which is a tubulin-targeted warhead.

Initial Phase I was done. Initial Phase IIs were done, actually led to the accelerated approval. But this was the confirmatory study that led to the worldwide approval of the drug. This enrolled patients with advanced urothelial carcinoma who had prior platinum chemotherapy and an immune checkpoint inhibitor, and randomized them to enfortumab vedotin versus the kind of old standard of care taxane chemotherapy, or what is available and approved in Europe, vinflunine, with the primary endpoint of overall survival. And I’m showing you slides presented by Tom Powles at ASCO GU 2021, also published in the New England Journal of Medicine. And these are the Kaplan-Meier curves led to the full approval of this drug in the United States as well as worldwide approval, an overall survival benefit.

Pretty much all the secondary endpoints were also met, which you can see below the Kaplan-Meier curve. PFS was superior. Overall response rate was superior as well as overall disease control rate.

Nothing is for free and I’m not talking finances, there are adverse events, but obviously, there’s also adverse events with the control group. You can look this up, I’m not going to go through the table in great detail, but the highlights I have pulled out on the right. So I’d say what I quote to patients, about half will develop a rash. Most of those happen to be Grade 1 or Grade 2. But there is Stevens-Johnson and there’s been some deaths occasionally. So this is something that we watch out for. Usually happens early, within the first weeks or months. So the way that this is administered is day 1, day 8, day 15. I make sure that either myself or my APP has actually seen the patients prior to their dose, to go over an AE assessment and assess for needs for dose hold or adjustment. Most often, dose hold is enough. But sometimes when it’s really severe, it’s dose hold and steroids.

Neuropathy, as not expected, based on MMAE, happens in about another half. You can see, because taxanes are there, it happened in a third in the control group, but this is one of the most common reasons leading to drug discontinuation. Because it actually works pretty well and controls the disease, but we stop it. A lot of times in the real-world, we stop it at Grade 2, which can start to interfere with quality of life. Also, like other drugs, this is something that tends to be cumulative.

Hyperglycemia was identified during Phase I and Phase II as actually lethal in a couple of patients. Not happening in the majority, but as part of the label you should assess a hemoglobin A1C coming in. Just because they have diabetes does not mean they can’t receive the drug, just make sure they’re glucose is optimally controlled and then start it. And then continue to monitor that. So those are things that, to me, stand out in terms of education.

So the EV-103 study was a study that looked at a lot of different combinations. I’m going to really talk about just 1 combination, but looked at a bunch of different combinations with the drug. The first combination to be assessed was enfortumab vedotin plus pembrolizumab, the anti-PD-1 drug, which you probably have all used across multiple tumor types. So there was an initial dose-finding, followed by an expansion, and this is the expansion data presented multiple times and then published earlier this year by Chris Hoimes and others, but you can see that this was first-line, those unfit, versus platinum. EV plus pembro, very nice waterfall plot. What I think is even more impressive is the spider plot to the bottom right where you can see it looks fairly durable, with an overall response rate of 73%, a CR rate of 15%. And as published earlier this year, those that responded, responded for more than 2 years on median. So a median duration of response of greater than 2 years.

That led to breakthrough designation and an additional cohort that was added, this was the same patient population, cisplatin-unfit, first-line, advanced urothelial carcinoma, was a randomized study. A randomized study, not to compare EV/pembro versus EV alone, but to really assess the contributions of each drug and also to get some data with EV in the first-line setting. And this is cisplatin unfit.

The EV dosing, in terms of the milligram amount, is the same, 1.25 mg/kg, but it’s given on a day 1 and day 8 q21-day schedule to match up with pembrolizumab.

And this was primary to look at overall response rate. again, these are not to compare each other. Really I would just focus on this middle column EV versus EV plus P, which is enfortumab vedotin plus pembrolizumab. And in a bigger cohort, really confirmed that this really does have a very high response rate versus historical controls, including some CRs. This is an earlier look at the data, so I would not say that the true progression-free survival or overall survival are really mature at all, but it’s nice to have that other data set. So this might lead to accelerated approval of this combination. We shall see.

This is while a Phase III — they’ve actually fully enrolled, but we’re awaiting for the data to compare head-to-head against platinum-based chemotherapy. So it’s possible that we’re replacing platinum chemotherapy for all-comers because that trial is all-comers, cisplatin for those that are platinum-eligible, and carbo for those that are not.

So this is the waterfall plot. This one happens to be broken down by PD-L1. There’s a very similar — actually it’s the reverse in terms of directionality, looking at nectin-4 expression. The bottom line is, it looks like PD-L1 doesn’t matter, nectin-4 on archival tissue doesn’t matter; it’s most can respond. You can see what’s nice is, on the left side of this waterfall plot there’s on 2 patients that had primary tumor growth. So, very good disease control.

So, shifting to another ADC and target, I suspect many more of you might know this because this drug was first approved in breast cancer, sacituzumab govitecan or SG. The target is Trop-2, so, overexpressed in a number of different epithelial tumor types. But when we look at triple-negative breast cancer and urothelial are two that very highly and fairly homogenously express Trop-2. This is different than most other ADCs where the linker is a little bit weaker, hydrolysable designed to release in the tumor cells as well near the tumor cells in the stroma. And because of that, the payload is a little bit weaker. So SN38 is the active metabolite of irinotecan. Actually has been studied by itself, so it’s something that can be given freely. It’s a little too toxic, but actually can be given freely. So a different kind of concept in terms of ADCs in general.

During the Phase I, my colleague Bishoy Faltas, we identified — he published the initial 6 patients in — with urothelial carcinoma in Phase I. and we said, hmm, this looks pretty good versus historical controls. This is the expansion n that basket study that looked interesting. And that led to the TROPHY-U-01 study.

So I’m going to discuss the main cohort as well as the second cohort. But what really led to accelerated approval is cohort 1. Which is of similar patient population to what was initially treated with enfortumab vedotin, metastatic urothelial carcinoma, post-platinum-based chemotherapy and immune checkpoint inhibitors. So that’s Cohort 1.

These who were enrolled in Cohort 1. You can see on the top right, median of 3 lines. So this was a median of fourth-line therapy. You can see that most had visceral disease. And a small subset interestingly had prior EV, which I’ll come back to.

So this was designed with the primary endpoint of overall response rate. And just like the initial accelerated approval of erdafitinib as well as EV, based on historical controls, this has a superior overall response rate and this led to the accelerated approval. So again, we’d expect 10%, 15%. This was powered to exclude 15% with the confidence intervals, which you can see there, 27.4% in terms of the overall response rate. Respectable, I’d say duration of response, because you can see the waterfall plot, the majority will have disease control, with the minority having disease progression.

I mentioned there were 10 patients that had prior EV. The hypothesis, and what was observed in Phase I, that some patients had EV that responded, different target, different toxin, should not have primary resistance necessarily. And the response rate with prior EV was basically the same, small subset so it’s hypothesis-generating. But patients can respond sequentially.

And this is the swimmers plot and spider plots. So, the PFS and OS is a single-arm trial, that’s why it’s not part of the label, but versus historical controls looks reasonably acceptable.

Not surprising, I think, to many of you who have used it for this diagnosis or for breast cancer, but we know from the drug development that myelosuppression is a dose-limiting toxicity. You can see that about a third had Grade 3 or 4 neutropenia, with 1 in 10 having febrile neutropenia. Prophylaxis with G-CSF was allowed. You can see that 30% received that. And then diarrhea, most often low grade, with 10% that were Grade 3. But Grade 2 diarrhea is really something. Might be 7 times a day. So what I recommend to all my patients is that we make sure that they have an antidiarrheal at hand and then take it if it starts.

You can see the other adverse events that are there. I know that in the breast cancer world, alopecia is a big one. It was seen in about half of these patients. There’s no notation of how many walked in without hair. So that’s why it may not be as a high as my breast cancer colleagues tell me it’s 100%.

So the confirmatory trial mirrors the EV-301 confirmatory trial. So same patient population, post-platinum and IO. The ADC versus physician’s choice of docetaxel, paclitaxel or vinflunine in Europe, with the primary endpoint of overall survival. This study is fully accrued. And we’re just waiting for the data.

So, like the EV-103 study, multi-arm combination study, what I’m going to go over next is the combination with a PD-1. This is a different patient population, so this is not in the front-line setting. This is those who have progressed after platinum-based chemotherapy patient population. Those are in line with the main pembrolizumab approval in advanced urothelial carcinoma.

So this was kind of hot off the press, very early data. so some of the patients had just been put on, but it was enough to make the late-breaking deadline. And as was presented at ASCO GU. 41 patients. You can see the majority had visceral disease. All had prior platinum. And some had a little bit more. Had a short response to prior platinum.

And this is the initial data with short-term follow-up. So this may change over time as patients get more scans, if they had a 29% reduction in tumor or not a confirmed 30% response, they wouldn’t count as a partial response. But a 34% overall response rate, again I only expect, versus historical controls. And we shall see. Again, the duration we don’t know because this is an early look at data, but at the minimum we would say that this appears to be tolerable.

You can see the different subsets here. I think this swimmers plot is interesting, but it is an early look at the data. So what I want to see, and I think what many of us want to see is, are we really making the immunotherapy work so we’re going to get these long-term responders or tail-of-the-curve? Hopefully. But we don’t know yet because this is an early look at the data.

This is just to point out that even those that had poorer prognostic grouping still had some responses.

We know that the individual adverse event profile of the components of SG, as well as pembrolizumab, this wasn’t clearly any worse because of that. Actually, let me take a step back. My caution with EV plus pembro is skin toxicity. We don’t have any true head-to-head data, but it might be a little bit higher. And then, which drug is it? Because we might treat it different. So with that regimen, I find myself maybe doing a skin biopsy if it happens.

Here, maybe there’s a little bit more diarrhea than we expect. We don’t really know. maybe there is more neutropenia, but we don’t really know that. Febrile neutropenia rate didn’t look so much worse. But anyway, just something that we would think about what could be cumulative, because we would certainly treat significant diarrhea differently if it’s immune-related versus just diarrhea with SG.

So to summarize just the 2 ADCs, EV with full approval in the post-platinum/post-IO space. Nowadays, that often might be second-line with platinum-based chemotherapy followed by avelumab maintenance.

There’s also accelerated approval for only 1 line of therapy. That was designed for the non-platinum candidates.

And the main AEs to watch out for are neuropathy and rash that happen in about half.

SG has accelerated approval, so we don’t have the randomized data yet, but does have accelerated approval in that post-platinum IO space.

Watch out for that third with high-grade neutropenia, with about 1 in 10 with febrile neutropenia. And two thirds with diarrhea, tend to be low grade, but can affect quality of life. And we certainly want to prevent them from getting to Grade 3. There’s a number of other interesting ADCs that are out there that I’m not going to go over right now.

Strategies for utilizing systemic therapies to treat patients with metastatic UBC

DR TAGAWA: Just kind of take step back, overall what I would say today and this may be changing as I alluded to with the EV-302 data coming, right now the standard of care is cisplatin-based chemotherapy, when we can give it. To remind you to go back and look at any cisplatin-based combination trials, the initial MVAC trials, gem-cis, etc. You always see this tail-of-the curve about 10%, that may be cured, tend to be mostly lymph node mets, but we can cure Stage IV with cisplatin-based chemotherapy. So that remains the standard until we prove otherwise.

We generally are going to use maintenance chemotherapy if they get cisplatin. But if they have recurrence after peri-operative chemotherapy, or something else, second-line also works.

We have 2 approved ADCs, EV as well as SGG. And we have a molecularly selected targeted therapy, erdafitinib, an FGFR-targeting. And just to remind everyone, that may be about 1 in 5. Some real-world data says maybe it’s a little less. But 15- to 20% of who we check. So if we don’t do any sequencing, and hopefully that sequencing will have the ability to look for fusions well, although it’s mostly mutations in the world of urothelial cancer, then no one can benefit from this. So, I think everyone should have tumor sequencing to see who is eligible for this. And we might see some other things that will open up possibilities for some clinical trials.