A roundtable discussion, moderated by Brian Rini, MD, discussed the latest data from the American Society of Clinical Oncology Genitourinary Cancers Symposium 2024 in clear and non-clear cell RCC. Dr. Rini was joined by David McDermott, MD; Elizabeth Plimack, MD; and Martin Voss, MD.
In the next segment of the roundtable series, the panel discussed important takeaways from RCC studies presented at ASCO GU.
—
Dr. Rini: I think we’ll move on to just a broader field, ASCO GU updates. Probably the biggest update there, as you’re all aware, was the updated overall survival data from the adjuvant pembrolizumab study.
There had been DFS benefit in the first 2 analyses and papers and a hint at OS with hazard ratios initially in the 0.5 but not significant in early, and now with what I believe was nearly 5 years of follow-up, there was a significant overall survival benefit hazard ratio, I believe 0.62, something like that, with pembrolizumab. Then the presentation before that was the other side of the coin with all the other studies, which have been negative. This was part B of the CheckMate study, which looked at single-agent nivolumab, ipilimumab/nivolumab, or placebo, which is a complement to the previously reported ipilimumab/nivolumab versus placebo study, which was negative, and that was similarly negative, one difference being a duration of 6 months of nivolumab monotherapy versus 12 months of pembrolizumab. But David, we’ll start with you. High-level thoughts. You obviously have melanoma experience with adjuvant immune therapy. Does the overall survival benefit of adjuvant pembrolizumab, how impressive is it and does it push you to use more of that drug in this setting?
Dr. McDermott: Overall survival benefit is not something we’ve seen in melanoma, even though we’ve had these drugs available first for stage 3 and now for stage 2 for several years. We’ve yet to see an OS signal, although we do see improvement in metastasis-free survival. Melanoma is a different entity with a different biology and a different recurrence pattern. For example, a lot of the recurrences in the melanoma trials are locoregional, and can be dealt with with a second node surgery, for example. The subsequent therapies are probably more active in melanoma than they are in kidney. For example, salvage PD-1 is active in those patients. CTLA-4 is active. BRAF MEK is much more active than VEGF. That may explain why we don’t see an OS signal there. But I think when you see one, I think 2 things are true: One, it makes the conversation with patients easier and in some ways, it makes it easier, in my mind, to recommend the approach.
I do think we’ll be using more of it, but I think in general, it was still left with 2 major problems. We don’t know who’s going to benefit and we don’t know which patients are going to recur. We’re probably over-treating still a lot of patients, so we need to insist on better selection tools than we have, also better management and reporting of toxicity. I think the last thing I would say is the questions that were asked at the end of the talk were really good and didn’t get answers, and they have to do with the impact of patients who recurred who didn’t get PD-1.
By my math, by the math of those people who were asking the questions, it was about half of the patients who recurred didn’t see a checkpoint blockade. We need to know how many of those patients contributed to the early deaths in that study? Are the people dying, the people who didn’t see PD-1 when they progressed? If so, then the OS signal is not as impressive in my view and would once again shape how I talk to patients. That information exists, and we need to know it, I think.
Dr. Rini: Yeah. Just for the audience, the questions, as David alluded to, one of the best questions I think Chris Ryan asked it. It was, to your point, David, how many of the patient deaths in the placebo arm got PD-1 or conversely didn’t get PD-1? If it’s 50/50, it’s very different than if it’s 90/10 or 10/90. As you say, the data certainly exists. If you have follow-up data and you have survival data, you have both. You just need to put them together. I agree. Impressive data, but the devil is in the detail as always. In this setting, it’s very different than the metastatic setting because you’re treating people who don’t have cancer by definition. Of course you’re over-treating. By definition you’re over-treating. I think it gives us a little different pause, at least for me, than it does in the advanced setting.
Dr. McDermott: Right. Particularly since we don’t see that long-term toxicity data. They report incidents and severity and then say toxicities were similar as reported earlier. In my mind, that’s not good enough. You’ve got to say how many of those side effects resolved? How many of them are life-altering? How many of them required immunosuppression long-term, all of those things? Because to me, I’m obviously an advocate for immune therapy, but when we’re treating patients who might not have cancer, I come at it the other way. I don’t want to over-treat patients, harm patients, because then what’s going to happen in the community is once you’ve harmed someone with adjuvant therapy, you’re less likely to go back at it again if you’re a community oncologist, I think, or any oncologist, and that’s going to limit the proper application of these drugs, I think, is the toxicity.
Dr. Rini: Yeah. I think it’s a good point. You don’t expect the numbers to change, because there’s probably not a lot of new toxicity years later, but to your point, it’s the persistence of and requirement for steroids and maybe long-term endocrinopathies, things like that. Those are the numbers that we really want to know.
Dr. Voss: I think it’s a different analysis that can be done than what we’re used to from the metastatic datasets because the treatment is finite. It’s like a 1-year therapy, so we can’t assume that it must be what we’re all used to seeing over time. I also think it’s not only about over-treating or harming the patients, which are obviously the big concerns we all have in our minds. It’s also about proper education and shared decision making. We have very smart patients and they are left to weigh the risks and benefits that we give them, but if we give them incomplete information, it’s hard for them. I think on the flip side, one thing that I actually found very helpful, we’ve all been obviously most keen on seeing the updated overall survival data, but I think it’s so helpful to see the updated DFS data for the placebo arm here.
As we talk to our patients, my own bias has always been that I think, and I talk to patients about this, that there’s a very clear tail of the curve, if you will, for patients after nephrectomy. We tell patients, the risk is the greatest after the first 2 years and then we space out the scans, it becomes less and less likely that your cancer will recur. If you look at the curves that Toni [Choueiri] showed during the meeting here and look at the placebo, the number of events that happened between year 2 and 3 and year 3 and 4 is just the same. There’s just as many patients that continue to recur at this point out.
They’re 50 months out and there’s still these events happening. When we used to tell patients the 2-year recurrence risk for you is probably somewhere in the 30% range, long-term it’s just becoming clearer where these things really fall. I’m looking forward to seeing this subgroup analyses. I’d really love to see the intermediate/high-risk placebo curve because that’s to me the most helpful one for discussing risks with patients. Those are most of the patients that we see in clinic. Right now, it still has all the M1 patients in there and so forth, but I think that’ll be helpful for the future.
Dr. Rini: Yeah, I agree. Betsy, other thoughts about the data?
Dr. Plimack: Yeah. Just to add to what Martin said about the control arm, there is a lot we can learn from that. I think it is recapitulating almost exactly the ASSURE control arm, which is as expected. We don’t expect biology to change as the number of therapies we have changes if we aren’t giving the therapy. I think when I saw that, I immediately start to translate this, how can I present this to a patient in clinic? Thinking it through, I think I would tell that patient there’s a 50% to 60% chance you’re cured right now and that anything we do is over-treatment. Conversely, there’s a 40% to 50% chance that there’s cancer somewhere. We can probably increase your chance of that cancer not coming back by about 5% with a year of adjuvant pembrolizumab, and that’s the basis on which I think we together can make a decision about that.
I think one of the other points that was raised during the Q&A at ASCO [GU] was this isn’t the best we have. We don’t give single-agent pembrolizumab to metastatic disease, we give combination therapies and is it best for that very high-risk patient to wait a beat, get that first or second scan when you see them recur, then give them something that has a stronger chance of controlling the disease and yielding a durable treatment-free survival or remission as David was saying? I think it’s philosophical and I don’t know that we have data to answer that question, but it’s definitely something I think about in clinic for those very high-risk patients who I guess it would be intuitive to think they need adjuvant, but maybe they’re exactly the patients will benefit most by waiting and exposing them to IO in a stronger combination.
Dr. Rini: Yeah. I think the same. Probably mostly for those M1 NED [no evidence of disease] patients. I hesitate for just that reason, Betsy, that I think they don’t need single-agent therapy, they need dual therapy of whatever sort you choose. Their recurrence rate is probably 78%, right? We just haven’t found it yet. Very few of those patients are cured. I think it’s only made the discussion more interesting, perhaps. Hopefully we’ll see some of the data and numbers that you all alluded to. Maybe we will, maybe we won’t, at least in the initial publication of this. But I think it’s good to ask these hard questions because we all sit in front of patients and we’ve had patients have bad outcomes toxicity wise and long-term insulin-requiring diabetes and all these things. Although they’re a little bit of anecdotes, per se, obviously, they stick with you and they clearly stick with that patient, and so we can’t ignore it.
Dr. McDermott: Right, and they add up. We often focus on those that are common, but it’s the rare ones that leave the biggest mark on both the patient and the clinician and you add them up, it gets, I don’t know, it’s 5%, 10%, maybe a little higher.
Dr. Rini: Yeah. The high-dose steroid requirement, I don’t know if they updated that. I think it was 8% from the prior presentation. But to your point, David, what I usually tell people is that 8 out of 10 tolerate it just fine, but 1 or 2 out 10 can have serious trouble, even lifelong. It depends on if you’re a glass half full or half empty person, I suppose. But I do think it puts in perspective that rare but serious toxicities exist and are super important in decision making.
Dr. Plimack: Yeah, and I think that’s one way to burn the IO bridge early in this adjuvant setting is if someone hits toxicity, but the other is at progression, would you go at it with a doublet? Would you do an ipilimumab/nivolumab or a TKI/IO combo and that patient progresses during the year of treatment? Or do we give up and say they’ve had their IO, it didn’t work, we’re moving on to single-agent TKI? I’m kind of uncomfortable with both scenarios. Adding and continuing an IO when it hasn’t worked feels a little bit wrong and expensive and maybe exposing them to more toxicity, at the same time saying you got one shot at IO, and it was single agent also feels wrong. I don’t know what people think about that.
Dr. Voss: For adding CTLA-4 to PD-1, we have data in the metastatic setting already. We know that salvage ipilimumab is not efficacious for many patients, for few. I think to go to ipilimumab/nivolumab then is exposing the patient to a higher risk of toxicity for a known response rate, which we there don’t have good data for how durable that response might be. Depending on what dataset you look at, probably somewhere in the 10% range. But no CRs reported on any of the trials that have looked at this in a sequencing fashion. I would not be so attracted to that knowing how ipilimumab can get you in trouble on other ends of things. I would not go personally to TKI/IO straight out of PD-1 failure given, again, what we’ve learned in the metastatic setting with CONTACT-03 where that strategy didn’t seem to add much, keeping the IO in place.
Dr. Rini: Yeah. I guess I’d say this is a big topic that we won’t have time to fully cover, but the biology of resistance and can you retreat and what if it’s 6 months or 2 years or progress? This is what we’re going to be talking about a lot. I think the survival data will only increase use, as I think David said. This adjuvant IO-resistant RCC population is going to be a growing one moving forward. You can bet we’ll start to see more and more trials in this space. CONTACT enrolled these people, but there was exactly 1 patient on each arm. TiNivo is I think enrolling them. There’ll be more. We’re going to start to see some prospective data sets and you would imagine you’ll start to see dedicated trials in this space because none of us know what to do and none of it is satisfying. I agree with you, Betsy. Single-agent TKI, in my opinion, is the very unexciting standard of care in this setting, but I don’t feel good about it. I feel bad about it. We need more drugs and more investigation.
Two last things to cover very quickly. One is the adjuvant nivolumab negative data, which is basically the same as ipilimumab/nivolumab and all those curves are overlapping. Just maybe real quickly, David, starting with you, how does that impact your thoughts about the pembrolizumab data and why do you think it was negative?
Dr. McDermott: To me, the pembrolizumab OS makes it easier not to talk about these other 3 trials personally, which I think is largely a waste of our time. I think the potential explanation on the nivolumab side is less duration of therapy, but who knows? The fact that ipilimumab didn’t add anything to nivolumab in that study was not at all a surprise to me. It’s not because I’m clairvoyant, but because we did that study in melanoma, and we did nivolumab versus nivolumab/ipilimumab and it was dead negative. Ipilimumab is adding more toxicity in that setting where you do not have a lot of tumor-specific immune cells. We’ve seen it before. Honestly, I think we need to move on to some of these other questions which are not going to be easy to answer.
We do have some answers to these questions in melanoma because in that melanoma KEYNOTE-054 study, which is the closest to the pembrolizumab study in kidney cancer, you could cross over to pembrolizumab within the study. We have a sense of what the activity is for placebo patients getting pembrolizumab and the response rate was near 40% with 14% CR, so very active. We also know that for patients who were on pembrolizumab and then progressed after 6 months off pembrolizumab, pembrolizumab was still active in some of those patients. Far less, I think the response rate was only about 11%, but that suggests that there’s some patients who might need chronic PD-1. Is that the case in kidney cancer? We need to understand that so we can get to some of these questions. It’s going to be hard to do without trials that look at that concept within the context of the trial, otherwise we’re just going to be doing all of these sort of retrospective analyses from large datasets, which it’s not as easy to understand responses in that setting.
630 Madison Ave.
Manalapan, NJ 07726
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.