The Uromigos Episode 134: PRISM Study—ESMO 2021

Dr. Naveed Vasudev discusses IPI dosing.

Episode Transcript

Tom:
I’m joined here by Nav Vasudev, who’s a colleague of mine in the UK and a friend. Maybe just summarize any conflicts of interest you have. You’re going to talk about the PRISM study, which is an oral presentation at ESMO this year.

Naveen Vasudev:
Yeah. Thanks very much, Tom. And thanks to you both for having me on. It’s an honor to join you. So yeah. Nav Vasudev, GU medical oncologist in Leeds, as Tom says. And delighted to talk to you about the PRISM study. You mentioned conflicts of interests. It was an investigator-initiated study funded by BMS, so probably worth clarifying that. So should I pick up …

Brian:
Yeah, tell us about the background and the design of the study.

Naveen Vasudev:
Yeah, yeah, sure. So, PRISM was a randomized phase 2 study, and essentially, we’re interested in the scheduling of the ipi component, the ipilimumab component in the ipi/nivo frontline combination in RCC. So those who are familiar with the combination will know the toxicity, particularly in those initial 12 weeks of treatment, can be significant. But based around that 3-weekly dosing of ipilimumab.

We’ve seen this new data from kidney cancer, but also, we’ve learned it from melanoma and lung cancer, that certainly the dose is relevant to the toxicity of ipi, but also potentially the frequency with which you deliver that ipilimumab. So, in lung cancer, they’ve played around with 3-weekly versus 6- and 12-weekly and going with the latter schedules based on tolerability. So that was the premise of PRISM.

So, the hypothesis was that if we gave the ipilimumab less often, we can make this combination more tolerable whilst maintaining efficacy. That was the hypothesis. So, we had about just under 200 patients randomized 2 to 1 in favor of a modified schedule in which we gave the ipi every 12 weeks for 4 doses versus a standard arm in which patients got conventional 3-weekly ipi and …

Brian:
The same dose of ipi, right? Still 1 per kilogram?

Naveen Vasudev:
Yeah, absolutely. Ipi 1, nivo 3. Yeah. That’s right. Right. So, we designed the study so that we could formally compare those 2 arms in terms of tolerability. So, the primary endpoint of the study was a safety endpoint that the proportion of patients getting a grade 3, 4 treatment-related adverse event. In terms of efficacy, because of course that’s key.

Are we making it more tolerable but less effective? So ideally, we would have liked to have powered this study for non-inferiority of the modified schedule, but look, this was a phase 2 study. That wasn’t going to be feasible. And at the time we sat down to design this, and I’ve got to acknowledge Tom has been key to getting this study done. So, I want to acknowledge that. Thanks Tom, but …

Tom:
That’s sweet, Nav. That’s sweet.

Naveen Vasudev:
At the time we sat down …

Brian:
He’s usually a hindrance. So, I’m glad to hear that.

Naveen Vasudev:
So sunitinib was still the standard of care when we were designing this, and CheckMate 214 hadn’t read out. So, we decided to sort of benchmark the activity of the modified ipi arm against the data that was available, and that was patients treated with frontline sunitinib. So, we based it around actually patients from the …

Brian:
Tell us the top line findings. You did see less toxicity.

Naveen Vasudev:
Yeah. Top line met its primary endpoint, Brian, so significantly less grade 3 or 4 adverse events in the modified arms, 33% versus 53% of patients with standard ipi. And then of course the interesting question is around, well, what are those adverse events then that are being sort of modified or lowered by giving 12 weekly ipi? And it was interesting.

It’s actually only sort of 3 or 4 AEs that in particular drive that difference that we saw. So, arthralgia in particular, colitis to some extent … to some extent. And actually, but the biggest difference was that in the rate of serum lipase increases, which arguably is not hugely maybe clinically significant, but it still speaks to the idea that we’re clearly modifying the AE profile.

Brian:
Did you have data on high-dose steroid use in the arms or hospitalizations or maybe other measures of difficulty for patients?

Naveen Vasudev:
Nah, would’ve love to claim to that, Brian. We were keen actually on getting these steroid uses. We don’t have that. I guess the other bit to this is to look at the percentage that we’re stopping treatment due to toxicity. And that was also … that was lower in the modified schedule arm. So, 23% versus 39% stopping due to toxicity. So, I guess that’s sort of another handle on that, but we don’t have that sort of data that you’re asking there.

Brian:
And do you think, in the standard arm, 39% stopping for toxicity is higher than CheckMate, right? That was 23% or something. Is that just maybe a reflection of not a big industry phase 3 study, or what do you think?

Naveen Vasudev:
Yeah. Yeah. You’re right. I thought about that as well. It’s quite a bit higher than what we saw. I’ve got maybe 2 things to that. One is, just to take it into the real world for a second, my instinct is that in real world it’s probably higher. So, I know this was a randomized … this was a trial setting, but whether there’s an issue around that.

And I don’t know whether we’ve learned a bit along … as we’ve gone along in terms of lowering our threshold to stop these drugs in terms of seeing that we can afford patients treatment-free intervals and that we see durability of response even in the absence of ongoing treatment. I don’t know.

Brian:
Go ahead, Tom.

Tom:
So Nav, question from me. When you looked at the efficacy parameters.

Naveen Vasudev:
Yeah.

Tom:
The first thing is how did the benchmark perform, the control arm perform against what we know from 214? How many patients in the control arm was it? I can’t remember.

Naveen Vasudev:
So, 64 in the control arm.

Tom:
How did they perform compared to benchmark 214? Benchmark … the 9-month response rate was 40%.

Naveen Vasudev:
Yeah. So, benchmark, if you go on the median, then in the … you’re saying the modified intense treat, you’re looking at 10-month median PFS. So yeah, we’re well within the confidence intervals of the 214 data. Similarly, if you look at objective response rates, you sat around that 40% mark comfortably.

So, I think in terms of those parameters we’re there or thereabouts to 214. I think the questions arise perhaps around, are we seeing the same durability of response that we saw in 214? So obviously, so this we’re reporting at a median of 20 months follow-up, so it’s pretty good.

But ideally, we’d love to see more follow-up with this data to see, are we going to see a similar plateau of the curve and where does that lie? And it’s probably we’ll say we didn’t see the same sort of CR rates and median duration of … we hit median duration of response with 19 months follow-up. So, we’re seeing median duration of response at 15, 17 months.

Tom:
… 2 arms, and the PFS in the 2 arms. Overall, they were very similar to one another. Is that fair?

Naveen Vasudev:
Yeah. You could not choose between them. The PFS curves completely overlie each other, Tom. So, there’s really nothing to choose between them. So, we can’t formally compare, as we’ve said, but informally, there’s nothing to choose between them.

Brian:
So Nav, I guess maybe the big picture question is, how much ipi do patients need and when? There are studies looking at obviously nivo versus ipi/nivo. There’s the OMNIVORE and HCRN studies, which tried to give it later. There’s your study of modified dosing. Do we know, melanoma has been through this a lot and we’re just starting, do we really know? Or what’s your opinion on how much ipi, and when and how frequent do patients need?

Naveen Vasudev:
Yeah, I think, if nothing else, what PRISM adds to this growing sort of story that we don’t know how best to give these drugs in combination. And what this, I think the key thing is that it really … it’s starting to highlight that we can maybe uncouple the efficacy of these drugs versus the associated toxicity, which is neat. Obviously if we can do that, but I don’t think we know the answers.

I think, like for instance, we’ve seen in melanoma some data, if you’re not really responding after 2 ipis, that another 2 isn’t going to rescue you. So small numbers, but I think it highlights that we don’t know the answers to those questions, Brian, but they’re important questions. And particularly as we think about using these drugs in the adjuvant setting, I think clearly becomes more pertinent still.

Tom:
Nav, was there any quality-of-life data?

Naveen Vasudev:
There was, Tom. Yeah. We used sort of the standard tools that are used in RCC trials like this. So, we use the FKSI-19 and EQ-5D, et cetera. The bottom line is, in terms of quality of life there was no time point was there a significant difference in quality of life between the 2 arms. What I would say though, is what you saw in the modified ipi arm, it didn’t really move from baseline.

So, it was pretty flat line. In the standard arm you saw this dip at around the 7- to 12-week mark, which kind of makes sense based on what we know. And then it recovered to baseline, and then actually slightly went above the modified ipi arm. Now it’s difficult to read too much into that. And as I say, they were never significantly different at any time point.

Brian:
That’s interesting. It seems different than the CheckMate 214 QOL data, which I don’t think dipped much if I remember correctly.

Tom:
No, it didn’t.

Brian:
In fact, it was sort of got better early on maybe because of steroid use, but … And I think this is, we just did a podcast with Janet Brown about her STAR study and kind of the same thing. These QOL instruments are pretty blunt tools to capture. Because you did show significantly less toxicity, but it wasn’t really captured in QOL.

Naveen Vasudev:
Yeah, that’s right, Brian. Yeah.

Tom:
Well, a couple of questions to both of you. I guess the first thing is that we’re giving ipi/nivo frequently to patients with metastatic kidney cancer. This is a randomized trial. It suggests you don’t need to give … the same. The toxicity might be a little bit better. Does this change practice for either of you, or do you think would you continue going with standard dosing? Brian, start with you.

Brian:
I don’t think it changes my approach to a standard patient walking in the door. I’ll still plan on sort of check my 214 dosing, but I think it does provide reassurance. And Nav mentioned some melanoma data that if you get to 2 or 3 cycles and the patients having significant toxicity, there’s nothing magic about 4 doses.

There’s nothing magic about every 3 weeks. I will often assess response if a patient sort of … if I’m worried about giving that extra dose, I’ll get some disease response data to put in the equation. So, I think we just need to be open to modifying, even if we’re not necessarily going to do it for every patient.

Tom:
And Brian, because it appeared that if you … I guess there are 2 questions, but I’ll try and be succinct if possible. It’s rather unusual.

Brian:
Pretty unlikely.

Tom:
So, would you consider giving a booster dose of ipi to any patients at any time after six-month period? Because if you believe that ipi is important, this data suggests that you can get away with giving boosters initially? What do you think about that statement?

Brian:
Yeah, I’m not sure about that honestly. I haven’t done that. I’m not sure about that. I’m not sure that’s true.

Tom:
So, you look at this data and this data tells you don’t need to give all four, but it doesn’t, it tells you the same … sorry. This data says you don’t need to force in all four. You should get all 4, if you can, but it doesn’t convince that you should give booster doses as time goes on?

Brian:
Correct for me.

Tom:
And Nav, where do you stand with it?

Naveen Vasudev:
Yeah. I’m with Brian. I think we …

Tom:
You can disagree with him by the way.

Naveen Vasudev:
Yeah. As much as I’d like to think that this should change standard of care, I don’t think it does. I don’t think it does at all. Three weeks will remain the standard of care, but I think it is important that it makes us think about …

Tom:
Brian, I’m worried that might be you in the background with your car.

Brian:
No, it’s not me. Sorry I’m outside. Apologize.

Tom:
Oh, now Brian, the OMNIVORE data showed that if you add ipi to nivo, and Mike Atkins did an excellent podcast on that this time last year, you increase your response rate by 20%. Life’s not all about disease control, and that’s suggesting that the addition of ipilimumab is associated with some disease control. So why not give a booster dose after a period of time?

Brian:
So, it was more like 10%. There were very few, if any, CRs. So, it’s not the same as giving it up front.

Tom:
Yeah, but that’s a 30% reduction in target lesions. I’m talking about a big … It’s not easy to achieve a response by just adding it ipi.

Brian:
Well, okay. But I’m just saying that a lot of those people never made it to salvage ipi. That’s the other thing you have to think about. If you start off with IO monotherapy and those patients progress and die, then you’ve done them a disservice. I’m not saying there isn’t a subset you can do that, but we don’t know how to identify that subset.

Tom:
But the addition of the drug a year into therapy, giving a booster dose to start with. I realize we’re a long way off pace so I’m asking provocative questions on purpose. It’s probably not doing any harm. And Mike’s data suggests some patients sustain some control.

Brian:
I don’t know where you’re getting it’s not doing any harm. Adding ipi will always add toxicity.

Tom:
Okay. I bought into that. I’m just pushing you a little bit if I may, because in lung cancer they have done studies looking at intermittent or different scheduling of ipi. Which again, hasn’t shown a big detriment to the efficacy.

So, my last question, I have to say, I’m not convinced that we’ve got the perfect regime, and had we started with ipi every 3 months rather than every 3 weeks we’d be probably having the discussion exactly the opposite way around at the moment, where you’d be saying well that [crosstalk 00:17:19].

Brian:
Agreed.

Tom:
And actually, so I think we need to be careful before we dismiss things. I agree the data is not robust enough to change practice, but I think the body of evidence does suggest that the addition of ipilimumab journey from an efficacy perspective is … probably adds some.

Question, the other opposite question if I may, the other hypothesis is ipi’s not doing anything at all. And it doesn’t matter when you give it because it’s got no activity. Just address that question for me.

Brian:
Well, there’s the 8Y8 study out there, which is nivo versus nivo/ipi. We haven’t seen results. I’m not sure when they’re coming, but that will address that. I think if you look at the IO monotherapy data, and then you look at the ipi/nivo, obviously not in the same study, you get the sense it’s adding to response rate and probably adding durability, but I’ll admit that’s cross-trial comparisons.

Tom:
Nav, what’s your take on ipilimumab? The single agent pembro data in Dave McDermott’s study showed response rate of 39% versus 40%. So, 39 for pembro, 40% for ipi/nivo. But the PFS was shorter. I think it was 5 months versus 9 months. Are you … ipi is required? And Brian, I’m going to come back to you for the last question. Do you think ipi is really needed for that long-term durability of response? Is this trial that you’ve done the right place to show that?

Naveen Vasudev:
So yeah. I think ultimately there are … We always talk about everybody in the same boat. There are a group of patients who I think are going to … who will do very well on single agent PD-1 inhibition, but the problem is we don’t know who they are.

And so, I think as you’ve been talking about, the idea of trying to rescue patients, I’m with Brian that doesn’t look robust as an approach. I think you need to get the ipi in initially. The debate will remain then clearly about how much more you then need to give and when you give it.

Brian:
I agree. I think it is important upfront for long-term durability, but I totally agree that dose and schedule and number of doses. And when do you stop in certain patients? We frankly have no idea.

Tom:
Last question. Well, second to last question. Brian, how important is more mature survival data for this trial? We’ve looked at … Nav, you’ve got some survival data at this point. It looks very similar. I think that’s fair. Brian, how much more would you … Because if you believe long-term ipi is important for long-term durability, you would expect that survival curve might be important in the future. How important is that curve?

Brian:
Well, I think it’s important, but I think this is just 1 piece of the puzzle and there’s probably a lot more trials to do to figure out how best to give ipi. Again, dose, schedule, and number of doses.

Tom:
Are we wasting our time asking these questions? Because it sounds to me like you’ve made your mind up. You’re going to keep going with the current schedule, whatever happens. What do we need to do? If a randomized phase 2 can’t change your mind, what do we need to do to change your mind?

Brian:
If you’re asking me, I think we should look and see what melanoma’s done and not reinvent the wheel. And a lot of the altered schedules with melanoma and with lung really haven’t been adopted. And so, I think it’s a good question we have to ask ourselves, is this the best place to spend our resources, our clinical research resources? I think it’s really important, but it’s just hard to study.

Tom:
[crosstalk 00:21:00] But my perspective on that for what it’s worth is exactly the same as what you said is that what we’ve shown is not very …

Brian:
We agree on this?

Tom:
Yes.

Brian:
That’s unusual.

Tom:
We do. We’ve spent quite a lot of time. Nav spent quite a lot of time showing that it doesn’t seem to make a huge difference doing it a different way. So, let’s not spend the next 20 years doing this in a randomized phase three. This is the point of the randomized phase two, a bit less toxic, no real difference in activity.

Probably easier to keep what we’re doing and look at other trials like COSMIC 313 and other bits and pieces rather than spending … refine this. Nav, what’s your take on what I just said?

Naveen Vasudev:
Yeah. Yeah. I totally agreed, Tom. This is not the right study to do to sort of maybe do a formal, big trial to show non-inferiority. You’ve seen the challenges with that. I think you’ve just been chatting to Janet, as you said about the STAR study and the challenges with that and trying to get a study like that done. And by the time you’ve done the standard, the world’s moved on again, potentially.

So yeah, huge challenges to doing these sorts of studies. But I think we all recognize the importance of them, but they are really difficult to deliver. And we’ve seen that already. But although we can see the relevance as you were kind of saying, we’re starting to add in other things to the ipi/nivo combination. So COSMIC 313 being an example of that.

Trying to reduce the toxicity seems relevant, but trying to deliver studies that are going to now change the standard approach is incredibly difficult.

Tom:
Nav, from my perspective, I think what this data adds, and I think, again, that it does suggest you don’t need to force in those first 4 cycles in the face of toxicity. And that’s quite important for the community, like either you or yourself suggest. From my perspective, I’m happy putting a significant delay in the third, or dare I say it, the fourth cycle of ipi.

I tend to give it if I can. And I’m very happy to put a six-to-eight-week delay in that. I don’t think you need to force them in on a 3-weekly time schedule if they don’t go in it’s do or die. My perspective on that is I think this data supports the fact that either delaying or not giving that is better than trying to force in those first 4 cycles. And remember it in the randomized trials, you really did have to comply with the protocol.

Otherwise, the patient came off protocol. So does actually add to that information. And congratulations on a terrific study, Nav. And I’m looking forward to seeing you very soon in Paris.

Naveen Vasudev:
Thanks, Tom.

Tom:
Keep well.