The Uromigos: Frontline Therapy in RCC

Brian: Welcome back. Our middle section here is on Frontline Systemic Therapy. We’ve titled it One vs Two vs Three Drugs for reasons that’ll become obvious. We’re going to start with the case, I’m going to back up so I can see, and I’m going to ask the panel to tell me how they’re going to treat this patient.

67 year old female presents with some symptomatic disease, fatigue, weight loss, cough. CT scan on the upper left shows this hypervascular renal mass, multiple pulmonary nodules that you see on the right mediastinal nodes and some questionable something in the liver. Gets a biopsy that reveals clear cell kidney cancer. Past medical history: hypertension on a medication. Has a remote TIA but no residual defects.

In the chart, there’s this chart lore of rheumatoid arthritis, but the patient doesn’t know much about it. She’s on low dose Prednisone but has never taken disease modifying drugs so it’s really sort of unclear what’s going on there. Has a ECOG performance status of one based on the symptoms, so mild anemia and has two risk factors, the anemia and presenting with metastatic disease.

This is a table we’ve put out that looks at One vs Two vs Three Drugs. Again, with all the huge limitations of cross trial comparisons. You’ve got David’s keynote 427 study as an example of IO monotherapy in this setting. Not the only example. We have obviously the two drug regimens; checkmate, two on four, and then the three IO/TKIs and then COSMIC, which will dive into a little bit more during the later part of this session with the triplet, and that’s listed there.

Let’s just go down, with this data in mind, Tian, I’ll have you start. How would you treat this patient? You can comment on the table data, you can comment on characteristics. They’re sitting in front of you. What’s going through your head as you’re selecting therapy?

Tian: Yeah, not uncommon presentation in our clinics. Definitely intermediate risk disease needs an IO based treatment. I would say that I would, probably with her history of rheumatoid arthritis, have a rheumatologist friend on call, but it doesn’t deter my enthusiasm for an immunotherapy based regimen for her. We’d be talking about IO/IO versus IO/TKI combinations. When we’re talking in these situations of de novo metastatic disease, I always caution a bit about aiming for a CR because those primary tumors don’t shrink to zero. But that is one of the benefits of thinking about Ipi/Nivo. If we have a great systemic tumor burden response, then maybe we go after the primary, and obviously we always think about our clinical trials as well.

Brian: This vague RA history doesn’t scare you away from Ipi?

Tian: It wouldn’t.

Brian: And you don’t want to do nephrectomy up front because there’s too much other extra renal disease, but you might consider it in a consolidative setting?

Tian: Yes.

Brian: What would you treat the patient with?

Tian: I think Ipi/Nivo would be my first choice.

Brian: Okay. Fair enough. Axel?

Axel: Yeah, I would have the same thoughts. First of all, I think metastatic load is not little, the primary tumor is actually quite small. Looking at the trace X data, you might consider that this is one of these rapid progressive types. You would like to have quick disease control. I’m not a medical oncologist, but these discussions are always around. Shall we use Pembro and Lenvatinib to have a quick response induction? But I’m also much more in favor of Ipi/Nivo in a sense because we’ve seen primary tumor responses. If patients respond, they have long and durable responses. And to my mind, if I look at the swimmer plots, it looks to me as that the time to best response is actually very similar with Ipi/Nivo and Pembro/Axi.

Brian: It’s always about three months, which is when the first came.

Axel: Yeah, exactly.

Brian: So it’s hard to know.

Axel: It’s more governed around side effects and what is possible. I would also consult a rheumatologist here. That’s it.

Brian: David. I think I might know what you’re going to say but go ahead.

David: I would agree with my colleagues that a pure IO regimen makes sense in this patient because the patient doesn’t seem to need an immediate response. I think the question of nephrectomy or not… Actually I think the COSMIC-313 data is actually helpful to that question, at least in my book, about whether to take the primary out or not. Ipi/Nivo performed in my opinion, obviously bias, pretty well in that trial. And that trial there was a higher percentage of patients were treated with their primary in place. I don’t think-

Brian: Although we don’t know response in one group versus the other.

David: No, I understand. But it made me feel better about doing less surgery. I don’t think surgery is wrong in this case.

Brian: She does have symptomatic disease. She has a cough. You said she doesn’t have bulky symptomatic disease, but kind of.

David: Well my bias is urgent. Those patients who need an urgent response, meaning a response in three months or they’re not going to be with us, get VEGF PD-1. Everybody else gets PD-1 alone-

Brian: How do you identify those who need an urgent response?

David: They’re in the hospital or about to-

Brian: They have to be in the hospital to get an IOT.

David: Or about to go in the hospital. Somebody who has a symptom, a pending complication. I have a pretty high bar for giving PD-1 VEGF.

Brian: Clearly.

David: But it’s not…

Brian: I’m trying just to figure out how high.

David: It’s high. It’s high. But the other thing I would say to what Tian was talking about, I mean I agree the RA is not a reason not to treat, but there’s not a ton of data on how these patients do with checkpoint blockade. And sometimes you flare their original rheumatologic problem and in this case that wouldn’t be a major threat because if the RA gets worse, well that becomes a symptomatic problem, but not necessarily a life threatening problem. There are autoimmune diseases where it can be a life threatening problem where I wouldn’t give checkpoint and some of these patients will get a second autoimmune condition, not necessarily worsening their RA but getting something else. That’s where the conversation of single agent PD-1 versus PD-1 CTLA-4 comes in. You might tell a patient that and if they’re concerned about the risk, just go with single agent PD-1.

Brian: I was going to ask you, maybe you just answered it. Who are you giving single agent PD-1 to?

David: To the wary. To patients unwilling to or concerned about the doubling of the toxicity, the serious toxicity being doubled by adding CTLA-4. If they’re concerned about that, if they’re not hyper motivated, I give them single agent PD-1.

Brian: How do you judge that? All patients are concerned about toxicity.

David: Well, there’s those who say “I walk through a wall” and others who are older and maybe not as…

Brian: They’re wary and maybe you are a little wary as well.

David: Yes. But also it comes from the melanoma world where some patients just get single agent PD-1 and if they don’t get a good response, you add CTLA-4. That’s what I find myself doing in the less motivated patients.

Brian: Even though all the trials I tried to do that weren’t really successful.

David: Yeah, well true. But they weren’t a car crash to…

Thomas: I thought we made up during the break phase.

Brian: Fair enough.

David: You should be giving it up for… I’m just, you asked me. It’s a small group of patients who get single agent.

Brian: It’d be a small group. Less than 5%.

David: Less than 20%.

Brian: Oh less. That’s not that small. Thomas.

Thomas: I think, I didn’t care what I say because it’d upset a lot of people. Look, these combinations are… It’s quite straightforward. If you want, the landmark overall survival of all of these trials are almost identical. The hazard ratio for survival of all of these curves versus Sunitinib for the intermediate and poorest patients are almost identical.

There are advantages of VEGF TKI/IO combinations and there are advantages of IO/IO combinations. The most important issue here is experience in giving the drugs the ability to give them well and because in a team that’s able to work together, the patients having well informed, being able to get steroids quickly, not pitching up to A&E departments and spending weeks on ward getting antibiotics. That’s the most important issue. We can have an academic discussion about the difference between IO/IO and VEGF TKI/IO.

Actually, if you’ve got a hundred people in the room and five years later down the road, you look at the five year survival, they’re the same for all of them. That’s quite an important place to start. There are a lot more similarities from differences between these. That’s number one.

David: He’s asking you to pick one.

Thomas: I understand that. And you folks have all picked if Ipi/Nivo and so I’m pushing back a little bit as I did.

David: So pick, pick one, pick one that’s not…

Thomas: As I did. Number one, there are the Ipi/Nivo data where there’s uncertainties about the role of Ipilimumab in this setting. We don’t know that, there is clarity around the role of VEGF TKI/IO in this setting. It’s clear that the response rate is higher when you add a second drug in the VEGF TKI/IO combinations have higher response rates than the single agent IO combinations, the PFS is clearly longer for that. And therefore there are clear advantages which have been demonstrated with the VEGF TKI/IO combinations. On the downside. I don’t think a lifetime of VEGF TKI/IO is particularly attractive. And I find single agent Nivo appealing.

If you said to me because the patient was symptomatic with cough and because they’ve got a history of rheumatoid arthritis, I would be giving them VEGF TKI/IO combinations. And that would be my preference. And I give much more VEGF TKI/IO than I do IO/IO for what it’s worth.

And then final thing is David’s now going to push me and say, which one of the three am I going to give? And I genuinely don’t mind and I’m not going to be pushed into that because on those three there are genuinely more similarities and differences. No, I can come up with an explanation for all three, I think access.

Brian: You’re going to give IO/TKI?

Thomas: Yes, I am.

Brian: It took us like, do we have any time left for this session?

David: That’s clear.

Brian: Dr. Cabo/Nivo, I mean Monty, what would you give?

Monty: Well, no surprise there. I’ll pick one. I’ll just say this is not a new debate. I must have heard three Uromigos podcasts on this by the way. We’re still talking about it. But I’m going to go with Cabo/Nivo here. I mean look, I’m just going here for the response rate of TKI plus IO. You know, I’m going here for this signal of improvement in quality of life we’re seeing with Cabo/Nivo. And it mirrors my clinical experience with it. I mean, heck, the strategy makes sense. We’re using a lower dose of Cabo in the context of 90 yard than we were used to second-line with Len-Pembro, we’re going higher. With Axi-Pembro, we’re using the same dose. I think there’s lots of good rationale for Cabo/Nivo here. I’ll stick to my guns as Dr. Cabo/Nivo.

Brian: Shocking.

David: You see that it wasn’t so hard what he did.

Tian: The problem is Tom is on all three of those papers.

Thomas: I wrote four papers actually. I actually, on Ipi/Nivo-

Brian: Conflict of interest are on the website.

David: Just so we know who said that and who didn’t say.

Monty: And I was on none of them.

Brian: Congratulations.

Monty: Thank you very much. Thank you guys.

Brian: That was summary of where we are. Obviously COSMIC-313 presented as mode this year. The next few slides are just some chosen slides including some from Monty’s excellent discussion. And I just wanted to tease out that data, our first opportunity to dig into that.

Monty: Could I just ask the friend Dave, you’d really offer that patient single agent Pembro.

David: I would recommend Ipi/Nivo if they were nervous about the tox, I would give them single agent PD-1. And I’d have the bottle of let’s say Cabo ready to go because here’s the thing, one of the biggest mistakes the kidney cancer field has made, and I was part of this, is treatment beyond progression. To me, that’s not a real thing in kidney cancer. Progression is almost always real, it’s not pseudo. You have to be ready to jump in with, if you can advocate for pure IO up front, you have to be ready to give VEGF at the first sign of progression. And that’s one of the weaknesses of some of these early trials.

Brian: That great. Agreed. Let’s turn to COSMIC now. This was presented a few weeks ago at ESMO. Stated this is the PFS analysis, this PITT population, which was the first 550 patients randomized. I think P stands for primary to get an early signal of a PFS. Tom, what’d you think of these curves when you saw them? Both the shape, the delta, the absolute numbers, et cetera.

Thomas: I was quietly pleased. I thought this is a nice curve. The curves go apart early. They stay apart. It’s a 25% reduction in the risk. I quite like these curves. Now if you said to me what would Cabo/Nivo have done had that been a control arm, that would’ve been a different curve I suspect. But do I like this curve? Yeah. Is it at the first time we’ve beaten a double. Yeah. Has the three previous panel members said this is our standard of care? Yeah.

I think it’s reasonable to say this is a really nice curve and it’s just made it a little bit nicer because we’ve beaten the standard of care of the previous three individuals.

Brian: Fair enough. David, I can’t tell whether you’re agreeing or disagreeing. I can’t.

Thomas: I think agreeing.

David: I know in the 313 data set, this is the best data that the investigators can point to. If this story remains the case over time, meaning the curves stay separate over time, then you can make an argument for using this triplet in a subset of patients. Now identifying that subset is going to be almost impossible but you could make an argument. Everything else is not so good.

Thomas: We haven’t got there yet.

Brian: We’re getting there.

David: Okay, fine. I look forward to the next one.

Monty: Spoiler alert.

Brian: Anything to add? Axel?

Axel: I would agree but spoiler alert.

Tian: Well the progression free survival was the primary endpoint of this trial and it met its primary endpoint.

Brian: Should it have been the primary endpoint?

Tian: Well, I mean it was a study designed with PFS as the goal.

Brian: I know. Should it have been?

Tian: Well to be honest, I think overall survival is a really important end point as well. Especially if you’re using all three drugs up front. And so, I’m waiting with bated breath for more OS outcomes from this trial. And I have some reservations. I don’t know, for example, if you’re thinking about a patient for Cabo/Nivo like Monty did, whether you would add Ipi to that patient, or if all three of us are thinking Ipi/Nivo for that particular patient, adding Cabo early probably would give a little bit earlier disease control. But we also know that with other VEGF/IO combinations.

Brian: Monty, anything to add?

Monty: No, that was well said.

Brian: This is the same population that PITT broken down by intermediate, which was the bulk of patients as you see. And then a relatively small sixty five-ish patient each porous group. I thought this was interesting. I mean you would think this regimen would do better in the worse patients, but those poorest curves are unusual, and caveats of small numbers. But David, do you want to comment on was this surprising to you and what do you think?

David: It went against the conventional wisdom, which shows why this was a good trial to do, because it helped shatter some of the assumptions we might have had about this idea. It was a good idea you needed to do this trial, but this result makes it hard to make the argument that some would’ve probably made that you need to be more aggressive in patients with a worse likely outcome. The data doesn’t support that; the reasons we don’t understand.

Brian: And porous is also-it’s anywhere from three to six risk factors. We don’t know exactly the breakdown and somebody with three and somebody with six are probably very different biologically. Monty did you want to…

Monty: I was just going to jump in with the comment here. When I got the slides for COSMIC-3, I only had eight days actually ahead of the meeting to generate my talk and I leaned on many folks in this room heavily for some input and advice. But what I’ll say is that I had this picture painted in my head before I got the data from the press release of who I give this regimen to. And it was that younger patient, very robust, but with multiple risk factors. And this just blew my mind. I just don’t know how exactly to reconcile this. It was tough for me.

Thomas: One of the ways I looked at this, the IMDC classification was designed around VEGF target therapy. That was where it was designed. And therefore in the poorest patients, they do the worst with VEGF target therapy. Therefore you might say when you add VEGF target therapy to the patients that do least well with that treatment, it shouldn’t come as a huge surprise the treatment hasn’t worked that well. And if you put it the opposite way around and say where does VEGF target therapy work best? It works best in good risk patients. And therefore you might say, actually there is an explanation for this, is we are giving the drug to the population that’s least likely to benefit from it.

Monty: Yeah, I would agree.

Brian: So would you have included favorable risk if you were designing this phase three? Because it was in essence TKI versus not. Same base therapy. If you want to show the greatest effect, you should include those patients, right?

Thomas: I would like to have done that. And I think there are ongoing trials that you are involved with, which are included in most those patients. And indeed there’s studies with HIF in that population. And I think that good risk population, which currently you could say Sunitinib is still the standard of care, I think in that good risk population, that’s a good place to test these drugs. The problem with this is then you can’t have Ipi/Nivo’s control on because that’s a problem…

Brian: Do you agree with that, David?

Thomas: Number one. Number two is that you have to wait a long time to get results in that population…

Brian: Okay.

Thomas: … and so this is a trial which I think was pragmatic. We went with a control arm which was reasonable, and from a global standard of care, intermediate and poor, and we are able to perform the trial in that manner. Now, are there subsequent trials that we could do differently and better? I suspect there probably are. Is this trial a positive, randomized phase three during the middle of a pandemic? Yeah, so it swings and roundabouts.

Brian: Yeah, fair enough. David?

David: To which point do you want me to respond?

Brian: Oh, pick your… I was asking you, do you think… Tom said Ipi/Nivo cannot be a control arm, including favorable risk patients I believe, and I saw your eyes roll.

David: Well no, I’m just remembering the days when that 214 was designed. Some people argued that we should include everybody for the primary endpoint, and those people lost the argument. If you look at the ITT population for that trial, it’s too bad. Because the benefit exists across IMDC types, which is not surprising, because as IMDC was designed in the VEGF era. There are plenty of responders in that good risk group, and you’re starting to see those curves starting to cross in the good risk patients, which is interesting. How the PD-1 VEGF is performing in that good risk population is also interesting, and potentially problematic for those people who advocate for those combinations.

Brian: Those people.

David: What’s interesting to me is not who’s right, but what’s the biology that’s driving those responses? It may not be what we suspect.

Thomas: Just to agree with David on that point for a change. Dave, so when you look at the Ipi/Nivo and long term data as it currently stands, the survival curves of that data, we’re beginning to see the Ipi/Nivo curves look okay. You look at the other curves which the VEGF-TKIs, which they’re looking the same. The current hazard ratio of full survival with the longest follow-up, Ipi/Nivo currently has the best in that population. The problem is the PFS curves, as you know, are pretty ugly at the start.

David: Right, okay.

Axel: But I think we also know from the emotion analysis that it’s overlapping. It’s not 100% congruent. But the patients who have more angiogenesis expression profile are favorable risk, and the poor risk are largely inflammatory driven. As you said, it wouldn’t surprise me that Cabozantinib doesn’t do a lot in the poor risk group. At the same token, I wouldn’t go for Ipi/Nivo in a favorable risk group, or for the combination because I think-

Brian: Because statistically, they’re less likely to be in that inflammatory group.

Axel: Yeah, exactly. Yeah.

David: I would just say we should stop focusing so much on early progression that doesn’t lead to death. Progression is important, and if it happens you should switch therapies. But OS is more important, long term outcomes are more important in a potentially curable disease.

Brian: I don’t think you’ll get any resistance on that. Tian, anything else you want to comment before we move?

Tian: People shooting for cures. I agree with Monty, actually. The patients with poor risk disease in my mind are the people who may need early disease control, and I was hoping this combination would have a really great effect in those patients. But unfortunately, we did not see that on cohort controlled populations.

Brian: Yeah.

Speaker: Brian?

Brian: Yeah?

Speaker: Very quickly, we have a question.

Audience Member:  I think it’s a great discussion, I agree with the point of biology, and statistically maybe poorest patients have different biology than good risk. My comment is from the statistical standpoint. I think you said, Brian, very small numbers. I want all of us to be cautious about over interpreting subsequent analysis which are underpowered, so just a point for the group.

Brian: Agreed.

Tian: You’re helping us with yesterday’s echoes, right?

Audience Member: Yes, what Matt Galsky said.

Tian: Subset analysis?

Audience Member: I echo Matt Galsky’s voice.

Brian: We only have the data at hand, and we’ll see how it matures, et cetera, et etc.

Audience Member: Of course.

Brian: We’re looking at patterns, right? We’re just looking at early patterns, so fair. Okay, this is the table of treatment exposure and discontinuation, then I have a graphic maybe that I stole from Monty on the next slide. But just to look at some of the raw numbers. If you look at the dose of Cabo, which remember, started at 40 in the combination like Cabo/Nivo. Basically, the median average daily dose was about half that, it was 23 milligrams. Number of doses of IPI, 58 versus 73, and those 15% of patients basically dropped down to get two doses. That’s where the 15% was made up. Then you could see the bottom treatment events, treatment related AE is leading to discontinuation. David, does doses of IPI matter? Is this a big problem if you only get two instead of four, potentially because you have Cabo on board that’s causing overlapping toxicity?

David: Well, this is one potential explanation for the result of the trial. That the investigators rightly were concerned about the toxicity of the triplet, so they wrote in criteria that limited the application of the triplet. In some ways, wrote in criteria that actually enhanced probably the activity of Nivo/Ipi. Because unlike 214, when you didn’t get four doses of IPI, you couldn’t get continuous Nivo. On 313, you could. In some ways, this is a very good trial for Nivo/Ipi, give the investigators credit. But adding Cabo to Ipi adds tox, which means doses of both agents being held, steroids given, and it’s essentially creating a regimen that is not practical. Or I shouldn’t say has limited practicality, particularly in the community, and explains part of the result. It’s a limited therapy, particularly in a poor risk patient, probably.

Brian: Yeah. Anybody else want to comment on the toxicity?

David: Because you’re holding the Cabo too.

Brian: Here’s a graphic. I did steal this form you, right, Monty? Yeah.

Monty: Oh, yeah.

Brian: Monty, maybe you can comment on this? This is your slide.

Monty: Yeah, this puts in a graphical form some of the data on the previous slide there on the left hand side. You can see the cumulative number of doses of Ipi rendered on the left over there. Proportionately four doses, which was 58% of the triplet arm. I put up there for historical comparison the data from 214, which is 79% of folks are getting all four doses. If you look on the right there, these are the proportions of patients getting more than 40 milligrams of prednisone, and it was actually double 58% the triplet versus 29%, which is the rate of use in 214 in the Ipi/Nivo with the arm. Interesting.

Brian: Is anybody worried about… Forget about just the toxicity to the patient, but the high dose steroid use just impacting the efficacy? Or is it just a drug delivery issue?

David: It’s both.

Brian: It’s both.

Monty: Yep.

Brian: Any other comments on the toxicity? This is Monty’s slide as well that looked at Waterfall. I thought this was your most interesting slide, Monty.

Monty: Well, thank you.

Brian: It looks at Waterfall, and makes the point that GA, you do see more tumor shrinkage. The response rate was marginally higher, but not as many deep responses. Again, within the limitations of the median fall, both 18 or 20 months or something like that. Is this a believable hypothesis? Is the adding Cabo somehow causing less deep responses? How would that work? Anybody? Thomas, what are you thinking?

Thomas: If you’re struggling to get the drugs in. Ultimately, I guess one of the issues with this is… We don’t know yet, and survival’s going to turn out to be important. The response rate and the CR rate in the triplet arm is perhaps less than we were expecting. The PFS is probably as good as we were expecting, I guess. Here, you might say if you’re struggling to get the drugs in, it might mean that you’re struggling to get those high CRs, those high PR rate. Because remember, we’re seeing 70% response rates with some of the VEGF-TKI/IO combinations, and in a triplet, you’d expect to do that, I guess at least. You wouldn’t expect it to be lower than that, and this does look a bit lower than 70%. The response rates are modest, the toxicity is a bit higher, and it may be that is dampening the ability to achieve these durable remissions. We don’t know that yet, and survival would turn out to be important, but that might be an explanation.

The other thing to say with this is actually neither arm had spectacular CR rates. I guess from my perspective, this is a demonstration that CR is not that great an endpoint. I actually would prefer if we moved the debate on from CR, because we started with 10% for Ipi/Nivo, we were told that it was higher than the 8% for Axi/Pembro. I thought that was a full argument. Then Len Pen comes out at 13 or 14%, and we then don’t talk about it anymore. Now we’ve got 3%, and this seems to again have been disappeared. I just think we need to let this discussion move on away from CR into more robust endpoints.

Tian: A couple thoughts. On the CR point, this trial had more de novo metastatic disease, like the patient you presented at the beginning of the session. To my point, those primaries don’t shrink to zero for resist criteria. They may be necrotic, but it’s very hard to tell that on a scan. That may factor in the difference in the lower number of CRs. Now, CRs were similar in the two cohorts, so it’s somewhat balanced. But higher numbers of de novo metastatic disease in this trial. Then I think to your point that we need to think about overall survival, and still get patients to live longer. Certainly more time will come, and hopefully we’ll see that from COSMIC.

Thomas: In the UK, we did a study called PRISM. PRISM was a randomized trial Ipi/Nivo versus Ipi/Nivo. One was spaced out, one was… it was NABS data from leads. Fantastic, investigating the shared trial and it was a huge effort in the individuals involved, which it’s terrific. That also showed very low CR rates of about 3%. I think we just need… We’ve got lots of reasons. Every time we get a study where we can’t explain the CR, we come up with lots of reasons. But sometimes, we just need to say maybe the CR rate is not that high, and there we are.

Monty: That’s true, agreed.

Brian: Monty, before we go to where systemic therapies go, you don’t mind, do you want to pull up some of the polls?

Monty: Yeah, yeah. If we could share with it.

Brian: If we could get the computer up, we have a few polls around triplet therapy.

Monty: Yeah. Yeah, I’ll actually start with the one in the bottom right really quickly over here, just to tell y’all that what Twitter does is it actually puts a little check mark by the right answer. For intermediate poorest disease, apparently Cabo/Nivo…

Brian: Or the answer that you gave.

Monty: …is the right answer. But at the very top over there…this was just a poll to get a finger on the pulse of what folks thought about triple therapy in general, and it seems as though the vast majority probably wouldn’t change current practice and offer triplets in the context of intermediate and poorest disease. I’ll say it’s a non-negligible proportion of individuals who might consider it, 25% with a pretty healthy number of responses there. I thought that was interesting too. Mixed takeaways, but for the most part, not a lot of folks voting for the triplet.

Brian: Okay. Was there another poll? I don’t remember if there’s another one.

Monty: I think that’s it for this section over here, yeah.

Brian: Okay.

Brian: Actually, let’s cover this first, because then we’re going to go with each faculty who provided a slide about what is frontline therapy going to look like in five years, and what’s the next trial you would do to achieve that? But before we do, maybe talk about this.

Monty: Yeah, so this is I think an interesting question, because it represents a lot of really great work that I think pretty much everybody here on this panel is a big component of. If you look to the top left over there, we have a schematic for the trial that Brian is looking at this novel triplet strategy with Len/Pembro, plus or minus a CTLA-4 inhibitor, or plus or minus a HIF-2 inhibitor. The right over there, you see this really, I think innovative approach that Tian’s pioneered through the Alliance Cooperative Group with Tony looking at Cabo added in based on response to Nivo/Ipi in the context of the pedigree trial.

Really, I love this design here from Bryan, Katie Beckerman, Scott, other folks here in the room today looking at this biomarker-based approach. We asked the audience, which one of these three do you think has the most chance of moving forward? It seems as though people are excited about biomarker-based therapy, followed by risk-adapted approach. I guess, Tian, your answer is the right one, it’s got a little check mark.

Tian: I’m very impressed.

Monty: Yeah. Finally, maybe a little less enthusiasm for novel triplets, but some still keen on that as well.

Brian: Just pause on this for a second. Tom, why would that Merck triplet trial be positive if COSMIC was, say, negative for OS?

Thomas: I think the Merck triplet trial has Len/Pen as the control arm. It has Len/Pen CTLA-4 as one of the arms, and Len/Pen HIF as a third arm. I think there was some data at Asset ESMO that looked at the combination of HIF and VEGF-TKI, and I thought the response rate, I think it was 57%, it was a bit higher than I expected. Actually, I thought that, “Okay, that looks quite interesting.” You put immune checkpoint inhibitor on that, and maybe that works. That’s a reasonable question to ask. Maybe it’s positive, weirder things have happened. Number one. I suspect it’s going to be good in the good risk patients, actually. I think that. Number one, so I like that arm.

Then the second arm, CTLA-4, PD-1, Len versus Len/Pembro. Len/Pembro’s going to be really hard to beat. Response rates are 77%, PFS in 20 months. That arm is going to be more challenging. The overall survival may be positive, because if we believe CTLA-4 has a role to play, it’s probably a long game and it’s probably those long term analysis, and those curves eventually going apart and staying apart, and that’s what we’d like to see. It’s possible that there are some patients who benefit through CTLA-4 who are not getting it with Len/Pen. The trial’s not blinded, I think that’s an advantage.

I think COSMIC-313 was blinded, lots of reasons and lots of advantages of having a blinded trial. But there are some disadvantages. One of the disadvantages of having a blinded trial with a triplet is that if for example you’ve got grade two toxicity with Ipi/Nivo placebo, you would be quite nervous. Grade two diarrhea for example, you’d want to give steroids and admit the patient. The blind made giving up a blinded triplet trial where you don’t know what your patients are getting, with immune therapy which is potentially life-threatening toxicity is complicated. I think there may have been more interventions because of the blind had it been an open-labeled trial. I think it could be positive, but I think Len/Pen is going to be hard.

Brian: So I want to move on to this. So basically we’re going to go down the row, I think. And I think, Tian, you’re first. And it’s vision for frontline therapy in five years, and your paired next slide after that is what you think the next trial should be. So we’ll do them both at once, and then we can talk about it. Try to do minute per slide. We’ll rocket through these.

Tian: Okay. Well again, I think patient selection is important, and a lot of this comes from your trial, Brian, where we’re thinking about transcriptomics gene signatures from emotion clustering, the angiogenetic biology, thinking about a TKI-based approach, maybe also double dialing on the HIF axis. So TKI belzutifan and then intermediate poor risk, more inflammatory myeloid gene signatures, IO doublets. So that’s where I think we’ll be in five years.

Brian: So, similar drugs, biomarker based. And then your next trial is…?

Tian: Gene expression profiling and doubling down on immunotherapies with the next new drug of choice depending on what that is versus angiogenesis looking at VEGFs and double downing on the HIF axis.

Brian: Love it. Let’s just go through and then maybe we can sort of have a group discussion. David.

David: So you need to predict standard therapy in five years. So this sort of builds on what Tian was talking about. Patients with an immune signature and their tumor, preexisting immune response. We should be pushing PD-1 based therapies. It doesn’t necessarily need to be PD-1 C two. Or there’s a lot of other interesting secondary checkpoint inhibitors like lag three, like TIGIT, like HHLA2 that could be explored in kidney cancer that we should be pushing for those patients because those are potentially curable patients. But what I’m focusing on here also is in HIF likely signature patients, we should be spending time as a feel to identify who responds to HIF. We’ve never had a great biomarker in kidney cancer but we’ve never really tried to come up with one in the right way and we’re now developing HIF inhibitors. It’s almost as if we’re taking a BRAF inhibitor and then treating unselected melanoma patients.

If you did that, people would say that doesn’t make any sense. That’s sort of what we’re doing with HIF inhibitors and we shouldn’t. This is a tumor specific marker. We should get serious biopsies in for metastatic lesions and identify those patients who are going to benefit because there’s no more tolerable approach in my opinion, than a HIF inhibitor for patients who are going to respond. But also you might be able to build on the HIF as Tian was saying, not just with VEGF, but with other potential ways of targeting HIF. Because you start seeing CRs in those combo trials and imagine seeing CRs with targeted therapy in kidney cancer. There’s a subset of patients who probably can get there with just targeted therapy. And finally, this goes back to emotion 151. If you’re not in either of those signature groups, you’ve got to go right to a clinical trial right up front. So those cohorts from emotion 151 that are not either of those two things, those patients should be going on trials right away.

Brian: I agree. I will say before you move on, I think as we’ve gotten optic up and running over the last 18 months, I think we’ve all learned a lot about how difficult the informatics are of determining cluster for individual patients. I mean, Scott, Katie and I have had a lot of sleepless nights over it and they’re not totally distinct buckets and it’s kind of where you dial up and down the expression level, et cetera. But I agree. All right. Your next trial is…? Basically as you’ve said.

David: Yes.

Brian: Biomarkers select. Anything else you want to add to this?

David: No, I think it’s doable and we’ve done some biomarker work correctly, but most of what we’ve done is been based on primary tumors, not metastatic lesions. Old samples, not recent samples. I mean we should copy from other specialties. For example, in lymphoma you would never treat a lymphoma patient based on a needle biopsy. So here we have this ridiculously heterogeneous tumor and we say, “All right, let’s stick a needle in it and decide what therapy is.” That makes no sense.

Brian: We dub that out.

David: It makes no sense. But I’m not asking people to do something that’s out of bounds for oncologists. We just should copy oncologists.

Brian: Mike Morris talked about lymphomatizing prostate cancer. That was the comment. All right. Axel?

Axel: Yeah, so I mean agree with the future. I thought I don’t add a slide to this. I’m also not an oncologist, but what I’m very interested in is basically we are seeing this group of patients, we could piggyback an arm like this into the existing trials like Probe and Nordic Sun that is investigating deferred nephrectomy. But we know from retrospective series that up to 11% of patients even in this primary metastatic setting develop complete response of metastatic sites. So we have an opportunity to defer these patients to no evidence of disease after having removed the primary. Whether it makes sense, yes or no is a different question, but at least it gives you the opportunity to stop therapy. And that could be investigated. Is it necessary to continue with Nivo for example, or accept. And then in the other arm you just follow up onto recurrence. The end point could be at least a two year treatment free survival.

Brian: I like it.

Axel: And 80 patients again should be doable because as I said, they are not many. That would be, for example of an 800 patient group would probably expect 80 patients to have these complete responses with Ipi/Nivo. I have to admit I haven’t seen them with the other.

Brian: Yeah, fair. All right, Monty, last but not least. Is this you? Yep.

Monty: Yeah, yeah, I know. It’s just shocking right that I’ve got Cabo/Nivo in there.

Brian: Double check the name.

Monty: But by 2027 I think we’re still going to be toying around with some of these principles of biomarker based therapy. What I’m really hoping we could do, and I won’t spend tons of time going over this, is really push some of these perhaps minimally toxic strategies forward into triplet regimens. And I like Dave’s notion of maybe introducing some other checkpoint inhibitors into the space. I threw out there – CBM 588, maybe other microbiome modulating approaches because I do think that’s a way to really avert some of these issues that we have with toxicity with triplets. And so the next slide I think really just sort of highlights the schema to get us there with a randomized study that looks at systemic therapy plus or minus CBM 588 and as Uromigos gains its listenership, all I need is 50 to a hundred billion dollars to do this trial. So if anybody-

Brian: You’ll have to talk to Mashup. I’ll have to talk to… I know there are efforts to actually do this trial. I don’t know if you can or willing to say or Tom or…

Monty: Tom, take it away.

Brian: It’s okay if you can’t say. I’m just curious.

Thomas: Well it would be an invest initiated trial. It would that be out of Europe. It would be a sort of a pan European affair. It would be a blinded trial. It would require… I think it would enroll well. And I think that from a European perspective we’d really like to try and do it. There are some issues around drug, of course. There’s issues around money and there’s issues around support and sponsorship and collaboration. But this is the trial that we’d really like to do.