Pedro Barata: Hello, everyone. Welcome to this UroToday video. I'm very excited today. We'll be talking about kidney cancer and maybe the first true biomarker for these patients. My name is Pedro Barata. I'm a GU oncologist out of Seidman Cancer Center, Case Western Reserve in Cleveland, Ohio. And I'm super happy, thrilled to be joined by Dr. Wenxin Xu. Great KOL in the GU world, is doing an amazing work on KIM-1, kidney cancer, assistant professor out of the Department of Medical Oncology at Dana-Farber and a friend and colleague. Wenxin, thank you so much for taking the time to be with us today.

Wenxin Xu: Pedro, I'm so glad to be here. Thanks for highlighting some of the stuff we're working on.

Pedro Barata: Absolutely. And first of all, congratulations because the conversation came out of your very, very elegant presentation at ESMO in Berlin around really strong correlatives for the COSMIC-313 trial around KIM-1. But I guess first things first, to level set Wenxin, do you mind just remind the audience why COSMIC-313 was put together and kind of quickly the highlights of the study, if you will? I know it's published, it's out there. Many people might have heard of it.

Wenxin Xu: Yeah, absolutely. So COSMIC-313 was a randomized phase three trial and the question the trial was asking was, can we build along the backbone of combination immunotherapy with first-line treatment ipilimumab and nivolumab in clear cell RCC? So the COSMIC-313 trial design, it was a randomized phase three and patients were randomized who had intermediate or poor risk RCC to either standard of care, ipilimumab and nivolumab with placebo or ipilimumab, nivolumab and cabozantinib, VEGF receptor TKI. And when this trial was done, the overall trial did show benefit where the triplet arm was superior for progression-free survival compared to standard of care ipilimumab and nivolumab. But there was no difference in overall survival and the combination arm was significantly more toxic. I think it's fair to say that despite being a positive trial, COSMIC-313 ultimately didn't change clinical practice much because of the lack of overall survival benefit and the increased toxicity. But the trial did provide us a really useful substrate where we have a combination arm with TKI and IO versus an IO-IO combination arm, which allows for some really great correlatives.

Pedro Barata: Well, fantastic summary. No, I agree with you. The first of many triplets we're going to be seeing testing. Right? The concept of combining different mechanisms of action, in this case dual IO, with a potent VEGF TKI like cabozantinib. Okay. So you did an amazing work around KIM-1. You have been involved in that area for quite a while now. Very helpful and forward-thinking regarding the utilization of KIM-1, so Kidney Injury Molecule-1. And the question here is how are you able to make it happen? Because right now it's not like we are already have it commercially available. I guess that effort is happening. We're trying to see where do we get it but we don't have it yet. So can you walk us a little bit about through how did you get that test up and going and how you were able to apply it to COSMIC-313 basically in the trial as the strong correlative work out of the trial that you present at ESMO?

Wenxin Xu: Yeah. Thanks, Pedro. So the background to KIM-1 is that it's been a molecule that as a medical field we've known about for a while. KIM-1, as you mentioned, stands for Kidney Injury Molecule-1 and has a bunch of different hats. This molecule has many roles in the immune system and in kidney biology. It's also called TIM-1. It's a structural homolog of TIM-3. So it's also an immune checkpoint. And it's also called HAVcr-1, it's the cellular receptor for the Hepatitis A virus. So there's really a lot of immune roles for this molecule and we've known for even several decades that KIM-1 is up-regulated when the proximal kidney tubular epithelium becomes de-differentiated. And that can happen in the setting of hypoxia, kidney injury. But it also happens in sort of these pseudo hypoxic differentiated states which happens when a kidney cell transforms into a tumor. And so in work that was initially led by Joe Bonventre's lab here at Harvard Medical School, we've known for a couple of decades that when a kidney proximal tubular cell transforms, it causes massive up-regulation of this molecule, KIM-1, which is expressed in the cell surface. And there is an extracellular domain of the KIM-1, which is cleaved by metalloproteinase 1 and 3 and circulates in the plasma. And that is what we're really taking advantage of because that is a detectable protein that can be found using standard electrochemical luminescence assays and can be used as a biomarker for kidney cancer. That also shows some aspect of immune response.

Pedro Barata: I see. Well super, super helpful. So how did you get access, right? To the samples from COSMIC-313? How did that process go? Kind of walk us through what made you do that or what made you use COSMIC-313 for that, for KIM-1 testing?

Wenxin Xu: Yeah. So the background for this is as you mentioned Pedro, we've been interested in KIM-1 as a kidney cancer biomarker for several years now. And this started with a series of other trials we previously analyzed, in the beginning really adjuvant trials, ASSURE, CheckMate 914. We also did IMmotion010 and several others and we found that across cohorts KIM-1 is a biomarker for detection of kidney cancer and for risk stratification of kidney cancer. And so the question came up, can this actually be used as an early response molecule for patients on immunotherapy? The preliminary data we got in this area was from CheckMate 009, one of the nivolumab studies that were done early on in refractory RCC where we really saw in a smaller cohort in this trial that a three-week drop in KIM-1 seemed to really correlate with long-term response to nivolumab immunotherapy PD-1 inhibition.

That was a small cohort. We validated this in CheckMate 214 where again we saw that among patients getting treated with ipil/nivo, a drop in KIM-1 before and after the first ipil/nivo dose seemed to really capture the long-term responders. And what was interesting was that this was really immunotherapy-specific in the first three weeks. The three-week change in sunitinib in CheckMate 214 was not prognostic for long-term response. We've since looked at that data and also JAVELIN Renal 101 and other trials. And so because of the CheckMate 214 data specifically, we really wanted to validate the ipil/nivo response and COSMIC-313 having ipil/nivo as the control arm was the ideal trial to have this done. And therefore we approached both BMS and Exelixis for getting these samples and it turned out to be a great fit because COSMIC-313 was one of the few trials where there were serial sample collections available at baseline. And after every dose of ipil/nivo there was plasma banked, which let us really get a long look at the longitudinal response in IO-TKI versus IO-IO for KIM-1.

Pedro Barata: Well, beautiful handling of the logistics, right? It's never easy to think. It's a lot of back and forth, I can only imagine and been there a few times and so kudos to make it happen. And also looking forward to the publication on this great presentation. I guess before let you go, Wenxin, you mentioned different settings where you can actually prospectively evaluate this, right? You mentioned the metastatic setting something like a treatment de-intensification or re-intensification is probably kind of what I took out of as you're walking us through what you found. But as you said, we could kind of think of KIM-1 as sometimes we are looking at circulating tumor DNA for other tumors such as urothelial carcinoma and think of like an adjuvant approach, kind of selecting the patients who are more likely or destined to do not as well with recurrences. So where's your mind in to what do you think is going to happen first in the kidney cancer arena, maybe both settings? Or is there any other setting where do you think the applicability of KIM-1 should be investigated sooner rather than later?

Wenxin Xu: Yeah. Pedro, I would love to see this get into the clinic. And exactly like you said, I think there are some provocative questions we can ask about how we can use this to make outcomes better for patients. I think just looking at the COSMIC data, what we found was that baseline KIM-1 was prognostic for outcomes, but also that the early change in KIM-1 and the change over time in KIM-1 was a really good biomarker for immunotherapy response. So I think one of the fastest ways to look at this prospectively will be in the metastatic setting with the clinically relevant question of can we identify who are going to be the ipil/nivo great responders? We know these patients exist, they get ipil/nivo, they do amazing. And in the one third of patients who have a 30% drop in KIM-1 after one dose of ipil/nivo by three weeks, you can identify those patients and they have an outstanding progression-free survival, overall survival, PFS over 70 months, OS over 85 months in CheckMate 214 in that one third of patients. But the opposite is true. The patients who don't have that KIM-1 response really don't do well. And I think prospectively a question to ask is whether we can identify those patients early because those may be those patients that we saw in PDIGREE that had early dropout high toxicity without benefit and didn't make it to step two, or in other words didn't complete their four cycles of ipil/nivo.

If we can catch those patients early, we can hopefully prevent early progression, prevent unnecessary toxicity and pivot them to a more effective metastatic therapy. And I think you also mentioned the adjuvant setting. We had an amazing presidential session at ESMO with IMvigor011 where we saw that ctDNA was able to identify which patients benefit from adjuvant atezolizumab immunotherapy. And I think a very similar question can be asked in kidney cancer. We now know that patients can benefit from adjuvant pembrolizumab, but we also know from your elegant discussion, I remember it very well, that a lot of patients are over-treated with adjuvant pembrolizumab. And probably if you really break it down, about 8% of patients, among all the patients we're giving adjuvant pembrolizumab to are the ones that really get the benefit. So could KIM-1 identify who those patients are, avoid over-treating everyone else and also potentially avoid under-treating the patients who are going to recur anyways on adjuvant pembrolizumab? We recently had a press release that belzutifan plus pembrolizumab met its DFS endpoint for adjuvant therapy at clear cell RCC. Can we identify who needs that extra HIF-2 intensification of their adjuvant therapy? And I think just like ctDNA has shown us in bladder cancer, maybe this is the way towards personalized individualized medicine.

Pedro Barata: Well, Wenxin, a perfect really well summarized. And I agree with you, it's really exciting. This is truly a beautiful example, story that really leverages the next steps. Right? You clearly understand how science evolves. Right? It's really easy to tell the story. You identify a marker long time ago, finally being able to test it, pre-clinical then in the clinical setting, then you go to the next step to validate in a prospective setting. I do think we'll see those trials with designs like the ones you just pointed out. Really smart, clever way to make that happen. It's such a pleasure to have you here, always learn from chatting with you. Super helpful and also again, congratulations for your great work and I'm sure we'll be talking again very soon.

Wenxin Xu: Always great to talk to you Pedro, and thanks again for highlighting our work. I'm really excited to see how the community takes this work and hopefully we can move it forward together.