Read the Full Video Transcript

Zach Klaassen: Hi, my name is Zach Klaassen. I'm a urological oncologist at the Georgia Cancer Center in Augusta, Georgia. I'm thrilled to be joined today for a UroToday discussion with Dr. Naomi Haas, who is a medical oncologist at the University of Pennsylvania. Naomi, thanks so much for joining us today.

Naomi Haas: My pleasure.

Zach Klaassen: So we're going to discuss the PROSPER trial, which was recently published in Lancet Oncology. This is one of the perioperative/adjuvant trials for high-risk RCC. Maybe just walk us through the genesis of this trial and how it came to be.

Naomi Haas: As you know, there have been a lot of neoadjuvant trials in kidney cancer that have been relatively small but have been of interest because of trying to shrink advanced, bigger, primary kidney cancers. And there have been many adjuvant trials as well. I was the study PI, study chair, for the ASSURE trial, which was adjuvant sorafenib and sunitinib versus placebo. All of the adjuvant VEGF-TKI trials were negative with the exception of the S-TRAC trial. With the advent of immune checkpoint inhibitor therapy showing a lot of activity in metastatic disease, there was great excitement about looking at immune checkpoint inhibitor therapies both in the adjuvant and the neoadjuvant settings in kidney cancer.

Zach Klaassen: Absolutely.

Naomi Haas: In the Eastern Cooperative Oncology Group, we had great interest in whether priming the immune system preoperatively would improve response rates in kidney cancer at high risk for recurrence. This was based on a couple of things, one of which was unpublished work at the time by Chuck Drake, who was looking at this in several mouse models, but also on neoadjuvant data in lung, melanoma, and breast cancer where small trials had indicated that there was a very good pathologic response for administering neoadjuvant therapy first. The thought was if you remove the kidney, you might not have enough metastatic or kidney cancer to prime the immune system. If the immune system was administered before the primary tumor was removed, that would be an opportunity to elicit a good response.

So with that, we designed a trial which ended up with a lot of negotiation. We weren't able to do a pure design where you would look at both neoadjuvant versus adjuvant therapy because the number of patients would've been huge, greater than a thousand patients. The way the trial was designed was that patients with a clinical stage T2 or higher disease were going to be randomly assigned to get either one dose of 480 milligrams of nivolumab before their surgery to remove their kidney tumor, followed by nine doses of 480 milligrams of nivolumab afterwards. The other group would be assigned to get straightforward surgery followed by surveillance, because surveillance at the time was one of the standards of care.

Because of the design, there were a couple of important things. One is we had to do a biopsy in advance to prove that there was kidney cancer. But this was also a trial that was jointly designed with the input of patient advocates. They felt very strongly that, if patients were having a biopsy which was not considered standard of care, that all patients should be enrolled. So the primary endpoint of this clinical trial was patients with any renal cell cancer regardless of histology. We enrolled both clear cell and non-clear cell, although we did have an additional primary endpoint, or a pre-specified endpoint, of patients with clear cell histology.

So we started off with everybody getting a biopsy, but that was changed so that at least the patients going on treatment needed to have a biopsy. That was the general design. It was 805 patients in the goal. As you can see below here, they had to have clear cell or non-clear cell histology, at least clinical stage T2 or higher. We did later on go back to allow patients who had oligometastatic disease that was planned to be treated to also be enrolled in this trial.

As you can see, the interim analysis presented at ESMO in 2022 by Mo Allaf showed that there was not a difference in recurrence-free survival, which was our primary endpoint, with a median follow-up of 16 months. The overall survival was not mature. The hazard ratio was 0.97 with a P value of 0.43. The trial was stopped a little bit early because it looked as if even following for another year or two was not going to change the outcome of the trial. The conditional power for primary and sensitivity analysis was less than 30%.

The primary paper was recently published in Lancet Oncology, and we did perform a pre-specified sensitivity analysis for recurrence-free survival where we censored the patients who did not have surgery or were not disease-free after surgery at day one instead of being counted as an event. You can see here that there was a little bit of a separation of the curves suggesting that perhaps the neoadjuvant arm was going to do a little bit better. But I do have to say that it did not reach statistical significance. The P value was 0.07 with a hazard ratio of 0.8.

What I want to emphasize though is that I think we're going to learn a lot from this trial. This is just a simple calendar that shows you that we did collect a lot of correlative tissue as well as blood specimens for patients participating in this trial, both in the group that was getting the neoadjuvant nivolumab and also patients who had standard care. So we hope to learn a lot from this.

These are the things in more detail that we were collecting. We're looking at some cardiovascular and quality of life instruments. There's a small study looking at bone density and the impact of IO on bone density. We've collected, we've banked all of the imaging for radiomics, and we're actively digitizing slides for a planned pathomics analysis. We did obtain some pharmacokinetics on the patients who were receiving nivolumab. We are currently analyzing tissue for some of the things listed here in the lab of Dr. Sabina Signoretti at the Dana-Farber Institute.

So the take-home messages that I have are really perioperative nivolumab did not improve recurrence-free survival in patients with renal cell at high risk for recurrence. But we did analyze very closely the adverse events, and they were consistent in incidence and severity with what we've seen in other adjuvant and metastatic nivolumab trials. I think there are a number of reasons that this trial was a little bit difficult. The lack of a placebo, I think you always question whether both physician and patient behavior might be influencing that. The lack of imaging immediately preoperatively made it a little hard to know whether the clinical stage was really the clinical stage. Perhaps, and we did find that there were a lot of patients who had a lower pathologic stage than I think might've been optimal for this trial.

There were a number of other things. The trial was conducted, as were all the other trials, during COVID. A lot of the patient visits were done virtually. Did that influence the behavior either more aggressively or less aggressively? But hopefully with these other analyses, we will learn a lot about how to design the next future neoadjuvant trials.

Zach Klaassen: Dr. Haas, thanks so much for walking us through that. Certainly, as you mentioned, the pathomics and radiomics and correlative analysis will be very helpful looking at some patients that may benefit from this treatment regimen. So the one thing I think, when we look at the disease space as a whole over the last 20 years, obviously we have the KEYNOTE-564 trial for adjuvant pembro for one year. We had DFS initially, now we have overall survival. And really, in the IO space, that's all we have so far. Is there a reason why perhaps this trial was negative? Is it the trial design? Is it the patient population? Is it the mechanism of drug? What are your thoughts in terms of why this was negative and maybe KEYNOTE-564 was a positive trial?

Naomi Haas: So I think it's a little premature to say it was the drug, because we certainly see nivolumab has a lot of activity in metastatic kidney cancer. I have lots of patients who've had CRs with nivolumab in the metastatic setting, even single agent. I think that there are a number of things. An adjuvant trial is always a cleaner trial. You know exactly what the patient's stage is, you know exactly who's really high-risk. And I think that was a challenge in the particular design that we had for this trial. I had said that before, I think having to analyze all comers, some of the other things I mentioned already, like the lack of the placebo design, the lack of imaging, but we have a lot of tools that we can use now that we couldn't even use then.

One of the tools that we can use is the imaging using girentuximab, which you could actually design a clinical trial with clear cell. We did look at core biopsies both for discordance, and we found good concordance between the clear cell diagnosis and what we saw in the renal tumor. But as you know, we're just looking at one core, and it can be very heterogeneous. So I think that would help, having some other imaging technology would help to sort that out, and certainly what we can learn with radiomics.

I think the staging is still going to be a real challenge. When you look at a regular CT scan image or MR, it's often easy to say, "Oh, maybe this got invasion of the renal pelvis." And then sometimes the tumor is actually smaller at the time of surgery, just even if you don't treat it ahead of time, the size can be somewhat different.

So I think there were a lot of reasons. I think there are a lot of things that we know now also about sorting out patient populations. There was some really exciting information at ASCO, for example, some of the KIM-1 data. We actually looked at KIM-1 in the ASSURE trial, and that was the first trial that showed that, if you measure KIM-1, it has prognostic value and it does follow according to change, to stage I should say. When they looked at KIM-1, even in the negative atezolizumab trial, they were able to show that patients who expressed high KIM-1 had a better outcome or benefited actually from atezolizumab. Whereas the ones that had low expression of KIM-1, it didn't seem to make much difference. I think that's another biomarker potentially that if it's prospectively validated could be useful. There are a number of other things that we have now also such as some of the expression of things like BAP1, SETD2, and some of the general expression by showing angiogenic stromal signatures and other signatures that have been validated in the EMOTION metastatic trial.

Zach Klaassen: That's a great answer. I think you hit on the improved imaging that's coming down the pipeline. I'm glad you mentioned KIM-1 too, because that was really interesting presented just a few months ago at ASCO. So I think just to wrap up, you mentioned this is going to inform trial design going forward. How do you see the adjuvant space in the next several years? And maybe even touching on what do we do in these patients that do get a good response and then recur down the road, how do we treat those patients?

Naomi Haas: So I think as far as planning other trials, there was a consensus meeting recently at the KidneyCAN conference. A lot of people got together, about 30 of us, just to go through different trial designs and stuff. I think there's still a lot of enthusiasm. I, for one, don't think we should rush right into doing a large trial right away without sorting out some of these issues. I think it would be really good to prospectively validate some of this stuff before we do that.

I also think it's going to be a real challenge trying to figure out, in people who've been exposed to adjuvant pembrolizumab, and more and more people will be exposed, whether those patients, when they recur, did they recur right away, did they recur a while later? Are they primarily resistant or not? I think that's a little bit more complicated than we're even thinking at the present time. There really aren't any trials that are designed at the present time that are specifically asking that question. Certainly industry and other trials are looking at recording how those patients do, but there isn't a trial that's been designed to address that at the present time, that I'm aware of.

Zach Klaassen: That was a great discussion. I really enjoyed your breakdown of the trial, and the Q&A afterward was excellent. Thank you again for your time and expertise, Dr. Haas. We really appreciate it.

Naomi Haas: Thanks and nice to meet you.

Zach Klaassen: My pleasure.