Martin Voss: Oh, it's my pleasure. Thanks for having me.
Zachary Klaassen: So you discussed three trials really within kidney cancer, but different disease spaces. So let's talk about the RADICAL trial. This was radium-223 plus cabozantinib in patients with bone metastatic RCC. Just give us a high level view of the results and maybe some of your thoughts on that.
Martin Voss: Sure. Yeah. So this was a cooperative group trial that ran for a long time. It was the first of its kind in many ways. It was set up to address an area of high unmet need and that is the treatment of symptomatic complications from osseous metastatic disease in renal cell cancer, which is something that we see quite frequently in our clinics. So patients in order to be considered for the study had to have metastatic renal cell cancer of any histology. They could be heavily pretreated or have no prior therapy. And they had to have at least one untreated bone metastasis on any type imaging modality. And if so, they could be considered for combination therapy addressing bony metastases, understanding that a standard of care in that setting is TKI therapy, in this case with cabozantinib. And cabozantinib was chosen in the design of the trial because it has had a long history of addressing osseous metastatic disease. There's data out of the METEOR study and from other trials that have looked at patients with bone mets specifically suggesting that cabo has an impact on bone turnover in a favorable way.
So patients were randomized on this trial to either get cabozantinib alone at a standard dose of 60 milligrams or the combination of cabozantinib plus radium-223. And radium in this case, of course, was chosen as a calcium emetic that once administered is integrated into bone matrix in again, areas of high bone turnover. So it essentially replaces calcium in the bone and delivers alpha radiotherapy to those sites of bone turnover. Patients were randomized one to one. And one of the important nuances was that on the radium combination arm, patients started with a lesser dose of cabozantinib and later could dose escalate to 60 milligrams if they tolerated things well.
And the primary endpoint for this trial, which is unusual, was symptomatic skeletal events. And with that, there were all kinds of means applied to optimize chances on both arms that patients would do well from a bone health and bone complication perspective. So for instance, everyone had to be on osteoclasts, the targeted therapies, unless there was a contraindication. And patients could be on opioids to treat bony metastases even at entry and so forth.
So the high level overview of the results is that the study was a negative trial. It actually closed early. It was meant to enroll about 60 patients per arm, but there was a pre-planned futility analysis that was conducted after around 90 patients had gone on study. There were numerically more events on the investigational arm. So the hazard ratio did not favor the combination and the study closed early. It's interesting when you then look at the skeletal related event data at the Kaplan Meier curves, but also at the overall survival curves that there seemed to be an early detriment on the combination arm, but those patients who stayed on later seemed to derive benefit and that the curves then cross and longer term there was the sense on the curves that there was benefit to some. And I think the question for the future is going to be what else can the study teach us about what radium does and who the patients might be who benefit from radioligand therapy to the bone as we think about other trials in the future.
Zachary Klaassen: And really a huge lift by the cooperative group to be able to do the first radioligand therapy trial in RCC.
Martin Voss: Oh, unbelievable. Yes. Yeah, yeah.
Zachary Klaassen: That's a great summary. Now the second trial, speaking of unmet need, non-clear cell RCC, first line, a Chinese study looking cadonilimab plus axitinib. Just give us some high level results from that and your thoughts.
Martin Voss: Yeah. So this was a straight-up phase one, two study. So two dose levels for the bispecific antibody were explored together with standard axitinib. And that bispecific that you mentioned is targeting two relevant targets, CTLA-4 and PD-1. And the first part was just dose escalation exploring safety and tolerability. And then the second part was a fairly small phase two single arm that was looking at one dose level and objective responses. And it enrolled patients, as you said, of non-clear cell histology. So that's an area of unmet need where we have seen good results with other TKI-IO combinations. So for me, the question was, was this going to look kind of like these data that we have for lenvatinib-pembrolizumab or cabozantinib-nivolumab, or does the addition of CTLA-4 as a target confer any added safety or efficacy signal?
Very quickly, the phase one part seemed very tolerated at least early on. No DLTs were seen. Both dose levels were explored with good safety and there was some efficacy observed also, but those were few patients, three patients per dose level. And then of interest, a second portion, so the phase two part, it's not very large, it's less than 50 patients, and it is sort of a smattering of different histologies including FH deficient translocation, chromophobe, even a few medulla and collecting ducts. And what that tells you is that the groups are really very, very small, so it's hard for us to be too certain about it. But broadly speaking, there was good efficacy seen across the board. The overall objective response across all of the patients was about 50%. A few CRs were seen. And the upfront failure rate was very low. I think there was only one patient who had progressive disease as best objective response.
So my impression here was... Oh, and then I should add the duration of response, and there's relatively short follow-up on the data set, was in the order of 18 months median. So to me, this sort of reaffirms the notion of giving TKI plus IO combinations in these patients across variant histologies. The only subgroup that stood out, the one patient who had progression was a medullary patient and there we always hesitate a little bit to consider IO-based strategies. In the frontline, these patients are typically treated with cytotoxic chemotherapy. But for the other variants, I think this shows us similar findings from what we've seen previously with other dual combinations. I did not get the sense with the limited data that we have that this was particularly different in terms of the efficacy. There was not a high rate of CRS by comparison. The duration of response is in the ballpark of what we've seen with cabo-niv or lenva-pembro. The whole gestalt of response seems very similar.
So I think it's helpful in that it again affirms with now a third combination that this has value for patients with non-clear histology. And I'm curious to see what the investigators, our colleagues in China are going to do with this combination. That bispecific has gone much further than other diseases, they actually have a label in GYN malignancy. So I'm curious to see if they're going to carry this forward further with kidney cancer.
Zachary Klaassen: Absolutely. Great summary of that trial. Last trial, the KEYNOTE-564. I always joke that ASCO officially starts when Dr. Choueiri hits the stage. And again, did a great presentation looking at ctDNA in RCC, which is sort of a new area that we're looking at. We've had a lot of information in bladder cancer. But just walk through the level or his results that he presented from 564.
Martin Voss: Yeah. So KEYNOTE-564 obviously is at this point well known by all because it's been presented many a time. There's been several publications out on this study itself, which was a randomized phase three trial to look at one year of pembrolizumab versus a year of placebo in patients with intermediate high or high-risk renal cell cancer after nephrectomy. And it is the same study population that was considered here for this exploratory analysis. Not all patients on the trial, of course, had samples banked away, but a good proportion of patients did, more than 300 per arm. And this is the first biomarker analysis that was seen off of KEYNOTE-564. So I think by virtue of that, it is super important.
So what was done? So effectively the investigators assessed two time points for peripheral blood MRD analysis and that was the baseline time, so around starting the pembrolizumab versus placebo, and then cycle five day one, which is about four months into therapy. And they did two types of analysis with these two time points. For one, they did several analyses around just baseline assessment categorizing patients into ctDNA positive or ctDNA negative. And then they also looked at longitudinal changes in the second set of analyses where they looked at baseline versus cycle five, what are the changes? Is someone positive turning negative, negative turning positive, or staying the way they were before?
So those are the two sets of analyses that Dr. Choueiri presented. It's relevant to understand, and he emphasized this during his presentation, that they are different iterations of the assay that was being used. So this is a collaboration with Natera. So this was the Signatera assay that was used to assess the patient's ctDNA status. So that's a tumor-informed assay. Essentially the way it works is you take a patient's nephrectomy specimen, you use, in the version that was presented here, whole exome sequencing to identify, in this case, 64 genes of interest at carry mutations. And they don't have to be biologically relevant mutations, they just have to be in the tumor. And then you use those 64 to generate a signature to manufacture PCR probes, and then you can use PCR technology for ultra deep sequencing in the peripheral blood to try to find that personalized 64 gene signature for the patient. And that is how Signatera works in the research setting and that is also by principle how it works in real practice because of a lot of providers are using Signatera in their practice already and can order it as a commercial test.
The devil's in the detail as always. There are different iterations there/. And obviously the company has worked through the years to optimize their assay. So there are different variations of the assay that can look at more genes and go greater depth and so forth. So the assay that was used here is sort of the second iteration of Signatera, which is the 64 gene exome. A lot of states now actually already have the 64 gene genome assay, which has considerably higher depth of detection. And then in the research setting, there's an even more sensitive assay that is now being developed. So with that backdrop, Dr. Choueiri looked at baseline first in his presentation and the baseline Signatera category being positive, being negative, associated significantly with the risk of recurrence. So patients as one might expect to where ctDNA positive had a much higher risk of recurrence. And that was seen both on the placebo arm and on the pembrolizumab treatment arm versus patients who were negative did comparatively well.
The point to make about the findings and the categorization of the patient was that the positivity rate at baseline was actually surprisingly low. Five to 10% of patients were positive. And we know from all we know about the placebo and the treatment arm on KEYNOTE that more than five to 10% of patients recurred. But the positive predictive value was extremely high. So the specificity, the reliance on a positive test signifying the risk of recurrence was extremely high, over 95% for specificity. And with that, patients who were baseline positive had an extremely high risk of recurrence.
In my review of the data, one thing that I've found interesting and I think useful in clinical practice is the fact that if you look at landmark recurrence risks at two years and you compare patients on the placebo versus the treatment arm together with their ctDNA status, it is seen that ctDNA negative patients stand to benefit notably less from the addition of pembrolizumab than ctDNA positive patient. In a patient who was ctDNA positive on the placebo arm versus a patient who was ctDNA positive and went on to do pembro, the risk of the two-year disease-free recurrence went down by an absolute of 20%, so dramatic differences. And 20%, as you know, is much more than what we typically quote patients as far as what to gain from adjuvant pembrolizumab based on the original data. And the odds of decreasing the risk of recurrence at two years in someone who's ctDNA negative, if you look at the placebo versus the treatment group, was in the range of seven to 8%, which is closer to what we typically I think we see.
So my takeaway from that just in terms of clinical application is that the sensitivity of the assay is not very high. And seeing a patient next week in my clinic who's ctDNA negative after nephrectomy is not super helpful for decision making because there's a lot of room for error there. But if a patient is ctDNA positive, not only is their risk of recurrence quite high, I also know that I can bring it down significantly with standard pembrolizumab, with the caveat that the assay that was being used in this analysis is not the assay that's on the market right now. And so that's where it gets really complicated.
The second round of analyses was looking at longitudinal changes. So going from screening to cycle five, so from day zero until like four months into placebo versus pembrolizumab. And there, and that's actually reassuring again to see as we learn more about ctDNA, pembrolizumab was much more likely to leave a patient who was starting off ctDNA positive to become ctDNA negative, so to clear ctDNA. That was seen in a good proportion of patients on the pembrolizumab arm and, as you might expect, in very few patients on the placebo arm. There were exceptions to both of these rules. And I think that again speaks to the imperfections of the assay that was used. And hopefully with technology rapidly evolving, that is going to get easier also.
And I think it leaves us with the question with longitudinal assessments as to how we can possibly escalate patients. So what if you were to see a patient on treatment and you actually see that they don't clear their DNA in the peripheral blood? Or what is it, what do you do if you have a patient who was negative and then turns positive? So I think we can now think about concepts for clinical trial design that allow for us to think through these things in a smart way and have adaptive designs where patients based on ctDNA trends go on to different treatment arms.
Zachary Klaassen: It's a great summary of the trial. You started off your presentation with the title of Is the Best yet to Come in Kidney Cancer?
Martin Voss: Yes.
Zachary Klaassen: So what would your quick summary be of that?
Martin Voss: The title is a little bit too simple. It's obvious that we're always going to say, yeah, the best is yet to come. But I thought it was in a funny way, it was like fitting to this because here are three very different topics and three topics with presentations that sort of point the way as to what might still be coming. I think for Dr. McKay's abstract, I think this was such a tour de force by the investigators doing what I don't think anyone anticipated could be done, looking at a clinically defined endpoint, which is very unusual on a randomized trial of a radioisotope. And I think it speaks to the fact that we can sort of think outside the box with smart clinical trial design. And I think it is a very fortuitous timing because we have several industry-led radioisotope efforts that are emerging in kidney cancer. And I do think that that opens possibilities hopefully also for patients specific with bony metastases.
Dr. Hu's data, the phase two non-clear cell trial, I think it was important that there was a non-clear cell abstract in the session. These patients do need more effort on our part in order to move things forward. And it was nice to see a novel molecule tested out in non-clear cell, not clear cell for once. So I do think there is more emphasis on that space. Then thirdly, I think ctDNA and other MRD tools, I didn't get to talk much about that during my discussion, but obviously there's not only ctDNA, there's also KIM-1 and methylome assessment from peripheral blood and so forth. I do think they're going to change the way we treat this disease and I think they're going to be particularly relevant in the adjuvant space where we need to escalate and deescalate appropriately for the right patients. So I think on all those ends, not only is the best yet to come, I think we are, I think, near term going to make great strides towards the betterment for our patients.
Zachary Klaassen: Congratulations on a great discussion and thank you for your time on UroToday.
Martin Voss: My pleasure. Thanks for having me.