Oliver Sartor: Hi, this is Dr. Oliver Sartor with UroToday we have a special guest from MD Anderson, Dr. Ana Aparicio, who's been looking at aggressive variants in prostate cancer for a number of years. Ana, welcome. Pleasure to have you here.

Ana Aparicio: Thank you.

Oliver Sartor: So let's start with a little bit of the past just to kind of catch our listeners up, and then we're going to pivot to the future. So let's talk about aggressive variant in your terms, some of the original definitions, some of the most recent definitions, and let's sort of lay that out for our listeners at the foundation. And then we're going to pivot forward. So let's start with a little bit of the past so we can get everybody up to speed.

Ana Aparicio: So the aggressive variant definition just basically came out of the need to be able to, from the clinic, parse out the observed heterogeneity. When you're in the clinic you can very clearly see that there are people who have very beautiful long-term responses to androgen signaling inhibition. And then there are those that don't, and we have no idea. We know this in retrospect, but we don't know how to tell upfront who's going to do what. And so to try to what we term the androgen-indifferent tumors to try to set them apart under the assumption that they had a distinct biology, which by the way is probably the biology that is most represented in our preclinical models, right? Because it's the bad tumors that grow when you biopsy and you put them in a mouse. Those are the bad tumors. I am actually of the belief that we have developed a lot of drugs that we're probably very good drugs, but we put them in phase three trials with all comers and the effect, because the androgen-indifferent were smaller subsets of the disease got diluted. So that was a motivation to try to separate this androgen-indifferent group from the rest of more androgen-driven disease. 

Of course, the poster child for that is a small cell or high-grade neuroendocrine carcinomas. It turns out small cells and high-grade neuroendocrine carcinomas, actually most of the histological variants as you know, sarcomatoid, squamous, but the small cells are the most common, have fairly atypical clinical features. So your garden variety prostate cancer will present with sort of neat round osteoblastic bone lesions and relatively small tumor masses and sort of proportional levels of PSA, if you will. Whereas the small cells or the histological variants will often have very bulky tumor masses, will often have lytic bone metastasis when they have lytic bone metastasis or even present with exclusive visceral disease. So this was an observation that happens in the clinic, the small cell cancers have these atypical features. Before my time, there was a phase two trial being run at MD Anderson. And whenever somebody showed up with this atypical clinical features, they said, aha, they have small cell cancer. We're going to biopsy it. The pathologist is going to tell us that they have small cell cancer. And then they're going to be able to go on our clinical trial, which is for small cell cancers only. And it turned out that 9 out of 10 times when people showed up with these atypical and virulent clinical features and we did a biopsy, the pathologist said, no, this is not small cell. We even found adenocarcinomas or just poorly differentiated carcinomas with or without neuroendocrine markers. And so that led to this idea that maybe, is that because the phenotype of the small cell cancer, of the bad cancer is unlinked from the morphology or because we have heterogeneity in the morphology and you just put your needle in the wrong place? We don't know. But the fact is this was a very clear observation.

And so that's where this idea that we could just use the clinical pathological features that are often associated with small cell cancer to select patients for trials that were targeting this androgen-indifferent biology. And so, the aggressive variant was just purely a framework to enrich for this androgen-indifferent biology and target it with certain drugs. And so we started with platinum-based chemotherapy, as you know, and we did show that if you just had the clinical pathological criteria, you did seem to benefit from adding platinum to standard taxane chemotherapy through a series of clinical trials. And we also found that in patient samples that they often had, this was many years ago, so we didn't have the sequencing that we have today, but it was early days of sequencing, and we found that a lot of those tumors had combined tumor suppressor defects. So in P53, RB or PTEN, so it had two out of the three. And the reason we had sort of confidence in this biomarker signature was really because of the published animal models. So at the time, various groups, this was circa 2005, 2006, so before the more recent 2017 P53, RB, PTEN models, there have been a series of animal models that have shown, if you knocked out P53 alone, nothing much happened. If you knocked out PTEN alone, I mean a little bit, but nothing much. If you took them both out now you had bulky tumor masses, you had poorly differentiated morphology on histological evaluation, and the same went for P53 and RB. So it was like, all right, well this makes sense, that these bad tumors would replicate what we're seeing in the tumors. So we've conducted a series of clinical trials using these aggressive clinical pathological criteria and the molecular profile, and it was very gratifying in a recent phase two trial, which was funded by the DOD where we added an anti-PD-1 to a chemotherapy backbone followed by a PARP inhibitor maintenance. And for those that reached randomization, the addition of the anti-PD-1 really almost doubled, not just their progression... Well, did double the progression-free survival, but prolonged the progression-free survival but the overall survival benefit even in such a small trial was great. It was substantial and quite meaningful.

The problem is that half of those patients didn't reach randomization. And so what we're now faced with is this heterogeneity within the aggressive variant, which shouldn't surprise us, because triple negative breast cancers, sure, they're different from the rest of breast cancers, but they don't all behave the same. Acute leukemias, they're separate from the chronic leukemias, but they don't all behave the same. So we expect some heterogeneity. And so we're exploring the reasons for that heterogeneity and trying to work on not just the epithelial component, obviously there's the whole genome sequencing, some genes show up, but nothing really robust. And so we're really trying to look a little bit more like everybody else, really, but at the tumor microenvironment and the phenotype of the tumor microenvironment landscape that combining those two features may give us a more robust biomarker to now select patients for distinct biological targets.

Oliver Sartor: Let me ask again, and I'm a little bit curious about this. Could you rephrase this discussion about trying to find a predictive biomarker for the utilization of platinum? We've been using platinum a long time. You've been using a platinum a long time. And what we see is some pretty phenomenal responses in a subset of patients. We actually published many years ago, went back and did some sequencing in the beginning of the genomic era and found, aha, BRCA2 was like magic. If you hit a BRCA2 mutant with a platinum-based agent, all of a sudden just beautiful things happen. But, there was so much more to the story. We knew that was just a small component. But let me phrase to you the question, would it be more productive to think about a predictive biomarker for platinum than defining the biology of the disease? And this would be a little more pragmatic. So I'm going to put that in your court and just kind see what you think.

Ana Aparicio: So yes and no in the sense that, I mean, that would definitely be one way to address the issue and would be very pragmatic, as you said. But I think what I worry about is that then you're benchmarking to a specific drug. And this is true for all tumors, but in particular for the aggressive variant, one of the pathways that has a barren DNA repair or replication stress, high levels of replication stress, and maybe those are the ones that are sensitive to platinums but also probably my bias is that the ADCs with the Topo1 inhibitors are going to be most effective in this subset of tumors, probably to other drugs that target that pathway. And so I worry that if we benchmark it to just one specific drug type rather than to the pathway, because resistance to drugs is such a sort of a complicated thing, we may go down a detail that just applies to that particular drug, but not to the biological pathway and miss out on, we may call something platinum resistant but that's not to say that those tumors may not benefit from targeting that replication stress perhaps with a better drug or what have you.

Oliver Sartor: So in a broader sense, and I'm kind of creating a big picture in my mind that may may not be accurate because I'm trying to understand your perspective, not mine, but it seems like we can sort of divide these tumors into the AR-sensitive tumors, the ones that have these beautiful responses to the ARPIs, and some go on for years and years. And then these ones that are very poor responders, which was a heterogeneous group, some of which are platinum sensitive and some of which are not. And if we follow the biology, what we're going to end up with is a group of aggressive variants, which the term that you've used for many years, and we're going to need to parse these out and to understand this heterogeneity in order to bring probably optimal therapies, some of which may be platinum, but some of which may be not. So if I were to put words in your mouth, and I'm not trying to, but I'm trying to understand the concepts, the idea would be to be able to look at this heterogeneous group and to really parse out the biology. The fact is that you got an amazing response in a subset of patients to the immunotherapy without necessarily the typical predictive markers of the MSI high is in fact a fascinating observation. I mean, that is a beautiful observation in and of itself. But then it begs the question, who the hell are these people and how do we get them to the right spot? So as we go forward, I can envision that the definitions of the aggressive variant, we'll remain a little bit broad, and the heterogeneity within that group will be the focus for the study because we need the right drug for the right patient at the right time in order to get optimal results. Is that a correct statement?

Ana Aparicio: Right. Right. Yeah.

Oliver Sartor: Okay. Okay, good. So let's kind of go forward. What's your next step? How do you intend to be able to move beyond the simple genomics, beyond the simple surfaceome... I mean, we have a lot of tools right now, but we all acknowledge that our tools are imperfect. So how do we get from our imperfect state today to a not perfect state tomorrow, but a more perfect state... that's a bad term. To a better state tomorrow from where we are today? How do we transition from point A to point B, what's the plan? That's a tough question, by the way.

Ana Aparicio: Okay. So my dream, and I've been thinking about this now for a couple of years, the execution is a little bit hard for various reasons, but my dream would be that we could create an aggressive variant platform trial, a la STAMPEDE. Have cabazi-carbo as your backbone, and then explore the various combinations and make those combinations tell us which are the patients that need to go this path. I mean, I still have four years out people on the C3NIRA trial, which is crazy, and they presented with awful, awful disease. And so if we could identify those people, and I'm hoping that a lot of the biomarkers that we're exploring in these trials will help us to, this platform trial would help us to then validate these biomarkers and say, yes, indeed. These are the people that are going to benefit from the addition of anti-PD-1, these are the people that need something else, and can we take them out quickly? Part of the challenge, obviously, and maybe I'm a little bit too detail-oriented, Chris Logothetis, who of course you know, because he's a big-picture guy, and I'm a very detail-oriented person, and so we have great discussions. One of the challenges with the molecular profile, the aggressive variant molecular profile, is that, sure, it makes sense biologically, but how do you measure it? Do you use next-generation sequencing? We were talking about this a little bit earlier.

Do you use next-generation sequencing? Do you use immunohistochemistry? And so this platform trial, I think would allow us to sort of validate and apply robustness to our biomarker profile in that way. And then the other question is that I struggle a little bit with honestly is do we stick to our cabazi-carbo backbone or do we pivot to some of the newer drugs? And then there's the challenge and sort of the constant, I always get on this soapbox of the pharmaceutical companies have to let us do combinations because we're not going to cure this disease with a single drug, and unless they figure out how to facilitate combinations, we're always going to be unsatisfied, and it's always going to be... That's really honestly part of the reason we stick to cabazi-carbo because we can get it and build on it. But for example, if I wanted... Right now I have another trial with adding an arginine depleting enzyme to the cabazi-carbo backbone because arginine pathways seem to be implicated in the resistance to this platinum backbone that we talked about. I'd love to add an anti-PD-1 to it because there's evidence that there may be synergy there. I can't get anybody to give me an anti-PD-1. And I think that's a real challenge in the field, we need combinations. And unless companies will allow us to combine their drugs, I just don't think we're going to get there.

Oliver Sartor: That's a tremendous challenge for this topic. We probably need to begin wrapping up, and I had like to ask you to sort of encapsulate the utilization of biomarkers as you go to define the heterogeneity next. And I know that's a big question, but I'm going to ask you to be relatively succinct. Which biomarkers, image-based biomarkers, immunochemistry, immunosuppressants, surfaceome, EVs, what do you think is going to lead us to the promised land or is it going to be some combination of all the above? How do we use these biomarkers we have today to help us choose wisely for therapies tomorrow?

Ana Aparicio: I actually think that it's going to be circulating biomarkers, and I think that the EVs are extremely promising. I think that we have a ways to go. A long ways to go. But I think ultimately that's really... Because the EVs capture more of the totality of the tumor. Otherwise, we're very focused on the epithelial centricity of the tumor. And so, I think that's the future. But I think a lot of work needs to be done to prove the specificity and that we're truly... Because there are many ways to interpret transcriptomic data, and you can always in some way, shape or form make it say kind of what you want it to say because it all depends on what genes you pick. I have this discussion all the time with the neuroendocrine markers. Okay, somebody defined for me what a neuroendocrine negative tumor is. I don't know because you all pick different genes to say it's a neuroendocrine tumor. So yeah, I think we have to be very rigorous and very disciplined about these markers. But I do think circulating biomarkers are going to be the way.

Oliver Sartor: Thank you. Ana, thank you for being with us today. I'd like to say that the field appreciates your focus on these very difficult to treat patients. We're all struggling together in the field to be able to optimize our therapy for these patients. We applaud you for taking a leadership role, focusing and trying to move us forward. So thank you, Ana, and thank you for being here today on UroToday.

Ana Aparicio: Thank you very much. As always, a pleasure.