Read the Full Video Transcript

Zachary Klaassen: Hi. My name is Zach Klaassen. I'm a Urologic Oncologist at the Georgia Cancer Center in Augusta, Georgia. I'm delighted to be joined on UroToday with Dr. Pedro Barata, who is a Medical Oncologist at Case Western Reserve University in Cleveland, Ohio. Pedro, thanks so much for joining us today.

Pedro Barata: I appreciate it, Zach. Thanks for having me.

Zachary Klaassen: So we're going to discuss a really important paper that was published earlier in 2024 in JCI that you were the lead author on, looking at transcription profiles of renal cell carcinoma. So maybe just by way of background, take our listeners through some of the important work that's been done in clear cell RCC for some of these transcriptomic signatures.

Pedro Barata: Yeah, no, fantastic question. So as you know, we're diving into genomics, and we started with DNA for the most part, I should say, limited benefit. We're doing that mainly for research purposes. We have not been able to nail down exactly what we're using at the genomic level to actually make decisions.

So we started with the DNA, with DNA alterations at the tissue level, as well as liquid biopsy level. And then we transitioned into the RNA-seq info. So these gene expression signatures have to do with RNA-seq data. So there's some provocative, I would say, and emerging data, first from the Atezo-Bev combo, which was one of the combos that, it turns out, didn't pan out. We do use a number of IO-based combos today for advanced clear cell RCC.

Actually, one of those combinations, Atezolizumab with Bevacizumab, did not pan out, but the studies that led to—well, actually developed the combo—the phase two and the phase three, the IMmotion 150, 151, we had a beautiful translational work looking at gene expression signatures for patients receiving TKI monotherapy, which was a control arm in those trials, Sunitinib or the investigational arm of Atezolizumab with Bevacizumab combination.

And so they came up with these gene expression signatures, or clusters, seven if you will—some more angiogenic, some immunogenic or inflammatory if you will, and then the others. And that's really promising data, very good data from Bob Motzer, Brian Rini, and others.

So the challenge has been to validate those gene expression signatures with other combinations. And we've seen data with Axitinib and Avelumab. We've seen data with Ipilimumab and Nivolumab combo. We've seen data with Axitinib and Pembrolizumab, and overall, we have not been successful at really validating the predictive role of those gene expression signatures to help us select the treatment that's going to be the winner.

So that's where we are. Very promising data. I think we need this last step to validate something that we can use in clinical practice. There are some ongoing important efforts. One effort is the OPTIC trial by the same author—Brian Rini is the PI. We also have that trial here at UH, and basically allocating treatments based on those gene expression signatures.

So let's just say it's a hot topic, Zach; it's not ready for prime time yet. It doesn't come up in the reports when you order all transcriptome sequencing data along with all genomic data, if you will. It is available to you for research questions and research projects. And we use that behind-the-scenes information that patients getting sequenced do have that data, to do this study that we put together in the JCI paper.

Zachary Klaassen: Excellent. And before we get into that paper, I just want to touch briefly on the importance of also looking at some of these transcriptomic signatures in the non-clear cell setting. We've seen some important trials over the last couple of years in the non-clear cell RCC. What's your thoughts on specifically looking at signatures in that population?

Pedro Barata: So notice that my answer was mainly clear cell. We almost have nothing in non-clear cell, to be precise. Part of the reason is, as you're probably going to touch base a little bit on that, non-clear cell is really a bag of different cancers and different diseases, biologically very, very different.

So it's very unlikely we would find something in common. So I would argue the difference is actually what's in common among all those. And so, I'm not really expecting to find something that will pan out, or be useful, in all non-clear cell subtypes. I think we have to probably look at papillary in a different way than we look at chromophobe, a different way than we look at collecting duct, translocation, etc.

So with that said, I think we're further away from finding a specific gene expression signature that would help us to treat a patient in a specific way. With that said, there are some exceptions to the rule. So we've seen some signals with specific therapies. I'm thinking mTOR inhibition, for example, in chromophobe; I can give you a few others. But in general, I think we're learning from the work being done in clear cell to try to apply it to non-clear cell. And so anyway, I'll stop there, but I think there's a big way to get us there.

Zachary Klaassen: For sure. And let's jump into your JCI paper. So tell us about the objective of this paper, and the study design, the population that you guys used.

Pedro Barata: Yeah, no, thank you for the opportunity to talk a little bit about it. And actually, you set the stage quite nicely. So when we put it together, we basically took advantage of accessing a large database in the United States by Caris. So we do test patients as part of standard of care; that data is available. And so we went after that data to understand what was there.

So we looked at over 600 patients—actually, most patients were clear cell; over 500 were clear cell—and then we had 149 or so, 150 patients with non-clear cell histologies. So very large cohort of patients with non-clear cell. Now, one of the limitations, as you can imagine, sometimes you send a primary tumor for sequencing, sometimes you send the metastatic site, whatever you have available, and we did not have clinical annotation. So that's one of the major limitations of that paper.

But the goal of the game was to understand: can we see patterns of gene sets, or in other words, what's the expression of these gene sets, or key molecular pathways, in these different subtypes? And is that something in common—how different or how alike they compare to clear cell? So that was really our main question, if you will. And so we really embarked on analysis of gene sets, and we used the clear cell as a comparator there, as a control, if you will. And then we have the analysis for a number of non-clear cell subtypes: papillary, mixed chromophobe, translocation, medullary collecting duct, papillary mixed with chromophobe—better numbers—a bit lower numbers, single digits for translocation, medullary collecting duct. And so that's what we wanted to do.

Zachary Klaassen: That's great. And so I know looking at your paper, there's a lot of results in that paper. Maybe highlight some of the key findings for our listeners that you found in your study.

Pedro Barata: Yeah, sure. So if we keep it simple, and we use that rose plot there, which is different colors representing the different gene sets. And what you can take from those pictures—I think the graphical abstract is probably the easiest way to look at the results—you can see that the clear cell line, the signal for the angiogenesis signature is overexpressed there. It is really present there. Whereas for the cell cycle, the fatty acid oxidation, and the pentose phosphate signaling, are increased scoring in the non-clear cell subtypes.

But when you look at the non-clear cell subtypes, you really see that's actually very different. So the message here is, biologically these tumors behave differently, and when we look at non-clear cell, we should really stop talking about non-clear cell and start talking about different cancers that they are. So that's that. There's a little bit more important data that I think, to me, can be helpful, because remember, we were able to get access to cell population microenvironment, if you will.

So you can have IO-based markers. So in that paper, you'll find a specific look into the patients with sarcomatoid/rhabdoid features. So we analyzed that group; we see the association with the cell population; we show the gene expression signatures along with DNA alterations of interest. And so, I would say not really surprising—it validates what we've seen before from the DNA alteration piece—but I think this is probably one of the first times—there's other efforts out there now—but one of the first times where we actually look at gene expression signatures across a multitude of non-clear cell subtypes. And that's really the effort.

I think folks will use this paper as a reference to support the idea of, "Let's look into papillary, let's look into translocation," etc., and look into that because you see the cell population along with DNA and RNA-seq data in there.

Zachary Klaassen: Excellent. That's a great summary. And I think let's look ahead. So let's take, as you mentioned, each non-clear cell, it's almost its own disease at the table. Let's take papillary, for example, given that it's the most common. How may these transcription and these markers, signatures, how may these inform future clinical trial design, and maybe eventually down the road, treatment decisions?

Pedro Barata: Yeah, so that's a great question. So right now, I don't think it's ready for prime time. We're not using that information to make decisions. Even if—which is a common question that sometimes comes up—if you forget the RNA-seq for a minute, just ask about sarcomatoid features. As you know, you can get sarcomatoid/rhabdoid features in any subtype. And so the question commonly comes up is, "Okay, what about a patient with papillary RCC and sarcomatoid features? Does that help me to decide adjuvant immunotherapy, for example?" And the answer is no, we don't really have good data.

Now, that said, we do see a signal in regards to immunotherapy and immune-based markers; we show that in the paper as well, that could make the point about maybe IO makes sense. And by the way, you look at data with pembro monotherapy—that's data from Dr. McDermott in that cohort B of his trial—you have data IO-TKI with CaboNivo or Lenvatinib-Pembrolizumab showing a promising signal, see 50% responses.

And I really think that's, for example, when we think we are in a point where TKI is a standard of care, it does make sense to actually, when you can, think about a combination of immunotherapy along with the TKI—you are looking for that signal. It doesn't work as much as in clear cell, but it's possible that some tumors will respond quite nicely.

Now, it needs to be validated in a phase three, randomized fashion that's ongoing. But that's probably why many of us in clinical practice, in the absence of level one data, are already considering an IO-based approach, based on the information I just alluded to.

So I think directly not ready for prime time, indirectly seeing signal around inflammatory component, if you will. I think it goes along with what we've seen clinically when we see responses to those systemic therapies that you consider in clinical practice.

Zachary Klaassen: And you touched on it a little bit in terms of validation, but where does this work go from your standpoint over the next, say, six to twelve months?

Pedro Barata: Yeah, I think the cleanest way to validate this work is really to design a biomarker-based study—clear cell that's ongoing. The challenge, as you know, Zach, in non-clear cell is finding those patients. It is far more difficult. There's also the logistical challenge, because you have to have tissue available, and then send that out, wait for the results, and get the results back to you, and make a decision based on that.

I would argue, probably won't see that before getting level one prospective data confirming or not that IO-TKI is the right strategy. By the way, we've just seen data with IO/IO presented at ESMO.

Where there is a signal there, response is 29%, was mainly papillary—more than half of the patients on OMNIVORE were papillary. And it was compared against Sunitinib almost 90% of the time. So it's really a lot of Ipi-Nivo for a lot of papillary tumors in there. And it's interesting, because we see the question before and after the trial data was presented, with an OS signal at 12 months, but no difference in median. Most people were not really convinced.

So you're not really doing dual-IO. So I think what's going to happen is, I think likely maybe IO-TKI is going to play a role. And then, I do think that we might use these RNA-seq data to see does everybody benefit from IO-TKI treatment intensification? Can we actually scale back? And for some of those individuals, consider monotherapy for IO—who knows—or not using an IO if the signal is not there from the biological point of view. I think those questions will pan out. We'll find them out, but we are probably a couple of years away from that.

Zachary Klaassen: Yeah, no, it's been a great high-level discussion, Pedro. I appreciate your time and congratulations on the work. Maybe just a couple of take-home messages for our listeners today.

Pedro Barata: Yeah, sure. We have gone from an anatomical description to looking under the microscope and how to treat people, to actually start going to genomics. We're not ready for prime time just yet; we're not using the genomic information—in our case, in our paper, RNA expression signatures—to make treatment decisions. But they seem to align with what we know thus far about the biology of these tumors.

When we think of non-clear cell, we really think about different cancers with different gene sets. And so, it's really important to understand these are different diseases, and I do think we need to pay attention to the trials that are ongoing. Hopefully we can evolve through these trials to get us to the next level, and get to the point where we're starting to do biomarker-based studies. We're not there yet. Maybe in the near future we'll be there.

Zachary Klaassen: Excellent. Well, thanks again, Pedro. Awesome discussion. And we'll make sure to include the link to your paper in the typeset for our recording. And thanks again for your time and expertise.

Pedro Barata: Absolutely. Thanks for having me. Always a pleasure.

Zachary Klaassen: Thanks.