Pedro Barata: Hello, my name is Pedro Barata. I'm a GU Medical Oncologist at Case Western Seidman Cancer Center in Cleveland. I'm very happy to be joined today by Dr. Kelly Fitzgerald. Kelly is a senior fellow at Sloan Kettering, has been doing amazing research in kidney cancer specifically. Welcome, Kelly. Thanks for joining us.

Kelly Fitzgerald: Thanks so much Pedro. Happy to be here. I appreciate the opportunity.

Pedro Barata: Absolutely. So, listen. Yeah. Congratulations. I mean, it's a big deal to have a paper in European Urology, right? And that's a big deal as a fellow never happened to me. So, very happy, you're brilliant investigating clinicians. So, congratulations.

Kelly Fitzgerald: I appreciate it.

Pedro Barata: Yeah. I'm bringing that up because we'll talk a little bit today on your paper that came out recently, right? This progression-free survival of the second line therapy for advanced clear cell CC in patients treated with IO based approaches, right?

Kelly Fitzgerald: Yes. Yes.

Pedro Barata: So, maybe we'll start there because, so you took basically a large cohort of patients, withed Sloan I believe, and what you did basically reporting on a couple of things. But one of the endpoints of interest was this PFS-2. That we've seen recently being reported in large trials like CLEAR, KEYNOTE-426. So, axi/pembro, lenv/pembro is an interesting endpoint, right? So, maybe I'll stop here, and I'll ask you, what was the motivation for you to do this in real world cohort? And why among the endpoints you chose, why PFS-2?

Kelly Fitzgerald: Thanks, Pedro. That's a great question. So, what we're trying to get at in this study is approaching this underlying dilemma we have of in the field of RCC management for first line, which is the therapeutic categories of ipilimumab with nivolumab, which we call IO/IO in the paper versus several combinations of TKI/IO regimens and really no clear guidance on how to sequence those. We also have TKIs available in the second line setting. So, what we really don't understand is for patients who get Ipi/Nivo or IO/IO and progress, can they get TKI/IO in the second line setting and actually catch up in terms of long-term outcomes compared with those who receive TKI/IO in the first line setting.

PFS-2 is an endpoint that can really help us assess this because it takes into account sequential lines of therapies. You start counting from the start of the first line of therapy and then you get to include outcomes on the second line and stop counting upon the second progression. So, it can help us answer this question of is there a difference in the way we sequence therapies and whether patients start out on IO/IO versus TKI/IO regimens is one of those winning out in the long term.

Pedro Barata: Yeah. I know that's a great answer and I mean, I completely agree. So, really well done with that design. So, I guess, you put together what over 170 patients or so. Interestingly, almost half or so got IO/IO, Ipi/Nivo and then the rest of your cohort got IO/TKI, right?

Kelly Fitzgerald: That's right.

Pedro Barata: And I'll let you walk through the major findings, although I have to tell you, we were just chatting offline, I have to tell you that it was quite interesting to me that actually you did not see a difference in the medium PFS-2 for these two groups of patients, right? Whereas one would predict, you're saving a TKI that you get usually get more response. And you did see that high responses in that group, but you didn't see actually a difference in PFS. So, can you walk us through the results of that study and perhaps offering an explanation for that if you will.

Kelly Fitzgerald: Certainly, so first of all, the designers, I think you mentioned this was a retrospective study where we identified all the patients treated at Sloan Kettering with contemporary IO combinations. So, either IO/IO or TKI/IO therapies. And about half of those, 90 of them received IO/IO, 83 received TKI/IO. In terms of the differences between the groups, I think it is worth mentioning those prior to talking about our results just because there were a few significant differences we found in the cohort. So, for those who received IO/IO, more of them had brain metastases or had intermediate and poor risk disease by either the MSKCC or IMDC risk criteria. And more patients in the TKI/IO group had undergone a prior nephrectomy. And so, these were significant differences and we wanted to account to those in account for those in our design.

And so, we had a few time to event outcomes. We were looking at PFS-2 and overall survival were the key ones. And so, as you mentioned, what we found is that there was no significant difference in PFS-2 between the two groups. And this was using restricted mean survival time or RMST that was adjusted for propensity score. The reason that we used this is we did want to account for the differences that were seen in the two groups. So that was why using a propensity score is important to us to apply for the modeling. RMST was used instead of the more traditionally used Cox proportional hazards model because the assumption of proportional hazards was violated in our data. And so, that was why we used this somewhat more on uncommon modeling approach to obtain PFS-2 estimates.

So that's a little background, but again to emphasize, when we looked at restricted mean survival time or RMST after 48 months for PFS-2 adjusted for propensity score. What we found is that it was 30 months for the IO/IO group and 33 months for the TKI/IO group, and that was not a significant difference. The P-value was 0.3 for that.

Pedro Barata: I see. So that's very, very interesting in my opinion, right? And also, of course you got fancy stats there, which are... I should highlight that because it really makes the paper stronger. And so, let me ask you, so we're talking about response rate in PFS, which are earlier endpoints, and then you bring up rightful cell, the perspective around overall survival. And going back to the question which I think is one of the IPOs of the study is, boy should we be changing our practice, what IO based combo should be offering IO/IO versus IO/TKI? And so far the answer could be probably not, probably we're making the right call.

But I guess one argument that the IO diehard people could make is, well yes, we're talking about medians and it's important to look at the tail of the curve. So, I guess all of this to ask you, are you seeing or do we need longer follow up, are you seeing any plateauing of the curve that for survival curves that suggests that even though the medians are no difference, there might be a difference regarding the durable responders or durable remissions favoring IO/IO versus IO/TKI for example, can you comment on that?

Kelly Fitzgerald: Yeah. No. That's a good question. And we do see a tale of the curve for IO/IO. That is something that we replicated in our real-world dataset, although it occurred at a slightly lower percentage than it did in the original CheckMate-214 study. But certainly, we were seen a tail of the curve, a flattening of the tail of the curve with the IO/IO group that wasn't necessarily replicated to the same degree in the TKI/IO group. So that effect is there, there's no question about it, but I think in the absence of having any biomarker that can help us identify the patients that will receive that benefit, that brings us back to the question underlying what drove this study, which is if we can't know in advance which patients are going to fall into that category. Are patients losing out by going on that regimen and potentially progressing, falling into the 60% or so of patients who have primary progression on Ipi/Nivo and don't do as well in the end.

Fortunately, what we found is that's not the case and that was the reasoning behind the study. What we found is that when we look at PFS-2 for patients who receive first line IO/IO versus TKI/IO, there was no significant difference in either PFS-2 or actually overall survival. So, encouraging that when we put patients on Ipi/Nivo and they progress, they can go onto a second line, receive the TKI in that second line and come out doing just as well.

Pedro Barata: Got it. That's very, very helpful. Kelly, maybe one last question, and it is also related to this, but I'm thinking about access to subsequent therapies, right? So, of course in a study conducted in academic center, as we see from your report, patients did get access to novel therapies and novel TKIs. I guess my question is, knowing that not every patient out there gets access to all the treatments that are currently available and approved, do you think that could impact the results if it was done a different real-world study? Right? Do you think that's the case? Can you comment on that?

Kelly Fitzgerald: Yeah. That's a great question. I think it's possible, but what part of what's at the core of this study is trying to understand is what we're doing for regimen selection for patients working and what we all use in practice is a clinical stratification system where we take into account various factors, patient's burden of disease, their sites of disease, comorbidities. And embedded in that are patient specific factors such as their preference and availability of different regimens. So, whether or not a patient has access to a certain TKI/IO regimen or Ipi/Nivo, I think it's at play here, in addition to all those other factors I just mentioned. And so, we all have this clinical stratification approach we employ, and one of the goals of this study was to understand is that working?

And I would argue based on our findings it is working and that that's something that's appropriate to continue doing. I think it's worth noting that in this cohort that we looked at only nine patients out of the 173 died prior to receiving their second lines therapy. So, the vast majority of patients were able to go on and receive a second line salvage if they had primary progression of disease with the first line of therapy. So, this is all evidence suggesting that clinical stratification is doing what it's supposed to be doing. Yes. Availability of therapies is embedded in that.

Pedro Barata: Got it. That's a great, great question and I think that a final piece of data that you shared with us regarding accessing, right? To subsequent lines of therapy is actually pretty relevant because the numbers from real world out there, large database, quite a quote, around 50% or so, patients getting second line, right? You lose patients from one line to the other. And we've been seeing over and over again in academic sites like high volume mechanic sites like Sloan, that number is probably higher, right? Of course, there's also there's a lot of factors including patient selection, et cetera. But it's interesting to see how we need to put the data into context, I guess as you brilliantly have done in this study. Kelly, thank you so much for being here and give us the... offering as the beyond your paper. Again, big deal, and it's a very important study. I enjoy reading it and I'm sure a lot of folks out there who will have the chance to read it will feel the same way. So with that, thank you so much for joining us and taking the time to chat with us today.

Kelly Fitzgerald: I appreciate that. Thanks so much, Pedro. Great to be here.

Pedro Barata: Thanks, talk soon.