Charles Ryan: Hello from APCCC 2019. I'm delighted to be joined by Professor Mark Rubin, who is Professor and Director of the Department for Biomedical Research at the University of Bern, not far from here, here in Switzerland. Thank you for joining us and we're going to talk about biomarkers today, and you have a presentation, and tell us what you're going to cover.

Mark Rubin: Well, I'm going to talk about the future of biomarkers in advanced prostate cancer. We've heard some great talks at this meeting about some of the struggles with developing predictive biomarkers, and that is biomarkers that predict therapy response. And so the problem is that they haven't been reproducible. And so I think the challenge in thinking about in the future, how do we improve that?

Charles Ryan: So are we moving to an era where we're looking towards liquid biomarkers, liquid genetic biomarkers so to speak, cell-free DNA, things like that? Is that where the pathology world is going in prostate cancer?

Mark Rubin: Well, I think I'm going to highlight three areas. I think that is absolutely one area, so the idea that because of prostate cancer heterogeneity, I think it's important that we try to capture all the information from the tumor and using liquid biopsies is one way for pathologists looking at specimens. We may need to also look at the microenvironment and this may help understand why patients may or may not respond to immunotherapy. But the third part that I'm going to talk about is I think we also need to start getting pathologists more involved in this activity. I think that pathologists have not traditionally been involved in evaluating samples from patients with advanced cancer, particularly prostate cancer.

Charles Ryan: Yeah. I mean there are a few instances where I have seen, my local pathologist has given me the standard report, the Gleason Score, et cetera, but I learn so much more from the tumor when I get it sequenced, for example, and the local pathologist isn't really involved in that process and isn't necessarily, even in their report, recommending that, for example.

Yesterday there was a nice presentation where we talk about MSI-High and it's a very big deal now clinically when you see an MSI-High tumor. And I didn't even know until yesterday that there's a strong association with Gleason 5, for example, in that, and I wonder if the pathology community is going to get to a point where they say, "This is a Gleason 5 tumor, it should be sequenced for MSI status, for example. That would help the clinicians who don't know that association exists.

Mark Rubin: Yeah. Well, I think we're, as in other areas in medicine, thinking about multidisciplinary approaches. And so you're right, traditionally it's outside of the realm of the pathologist to order those genomic tests and I think that needs to change, but it has to be part of a partnership with oncology, and in Europe with urology.

Charles Ryan: Yeah. I also think, as a clinician and somebody who focuses on advanced disease and somebody who participated in the Dream Team, Stand up to Cancer efforts in California, that we're going to need to get to a point where we're doing more metastatic biopsies and that's going to be I think, require a partnership in how we analyze those samples. That's something that I think the community more or less needs to get on-board with. It's not just the pathologists and the oncologists, it's also the interventional radiologist, for example, and those who are acquiring the tissue. Do you see that happening?

Mark Rubin: Absolutely. So in the precision medicine programs I've been involved in, not just for prostate cancer, but for other advanced cancers, we found that extremely important because very often the tumor that we see before the patient is treated looks very different. And it's not that it's a different tumor, it's just that only a select portion of that tumor goes on to progress through mechanisms of resistance. So we see that very commonly in prostate cancer.

Charles Ryan: I want to ask you a couple of specific questions about some of the biomarkers that I think are causing the most confusion and are common. One is TP53. When we get sequencing results back, certainly when I do and I see p53 loss in the tumor, I know that I'm dealing with a more aggressive cancer. I don't know what to do about it but, I kind of stop there, and I know when I'm facing that patient that we're dealing with a more aggressive situation. But what should sort of the typical clinician know about, for example, p53 and what it means in prostate cancer?

Mark Rubin: Well, I think you brought up a great example, whether it's p53 or RB1 loss, PTEN loss, these are indicators of a more aggressive disease. Any two of them would really tell you that patient is very likely to have a more aggressive disease. How we actually use it in clinical practice is something that I think that's part of the future. I think we have to establish the trials to really rigorously say patients should be treated differently, more aggressively, but certainly, if a patient is being considered for active surveillance and they have these alterations, one would take note and we need the hard evidence to really say they should be treated differently.

Charles Ryan: I mean, I wonder, there was a very nice paper, Wassim Abida, I think you're on that paper.

Mark Rubin: Yeah.

Charles Ryan: Maybe you're the senior author, where RB loss is probably the most important biologic factor emerging post abiraterone, for example. And I wonder if we will look at a situation where in the future we may look for RB1 loss and say, "I'm not going to treat this patient with a secondary hormonal therapy like abiraterone or enzalutamide because that resistance factor may already be there." It hasn't obviously, passed the muster of full-on development as a biomarker, but again, we know that that tumor is a more aggressive one.

Mark Rubin: Yeah, I think this concept of a predictive biomarker where they're selling it to treat or not to treat a patient is very important. The concept that you would not treat somebody because they have a mutation is dangerous, and we see this particularly with AR. So in AR-negative tumors, one might say they shouldn't have the AR predictive therapy, but in fact, they may benefit and we don't know that for sure, because they still have AR signaling going on. So, we really have to be cautious about how this then translates into clinical care.

Charles Ryan: I've had a few conversations recently about treatment-emergent disease and RB1 loss is a great example, where if we were to look at biomarker progression, biomarker evolution, so to speak, and to demonstrate that after, for example, an abiraterone-based therapy, we have not lost RB1 but that would be a better prognostic indicator. And that may be something we may want to consider for example, in the stratification of trials. And I think that's the third place where we need to be thinking about these biomarkers, which is instead of necessarily prognosis and prediction is stratification.

Because I wonder in all the clinical trials we've done with narrow margins of benefit, if we were to go back and look at the p53 intact tumors versus those that have lost it, if we see a different differential outcome and maybe an imbalance in populations based on that. As a clinical trialist, I wonder what it would take for us to get to that point.

Mark Rubin: Yeah, I think this is one of the areas that we're really struggling with. I would recommend that any trial have some element of genomic testing in it because I think we need to have this hard data upfront to then potentially retrospectively evaluate it for the next trial. And unfortunately, I think we've seen so many different platforms used and some assays that are not reproducible, and this has created a lot of questions in the whole field. So I think as a group, and this sort of gets back to my observations about pathology, I think pathology needs to get more involved in this as well.

Charles Ryan: I think that's a challenge that you just identified, which is one lab telling us there's RB1 loss, or one assay I should say, may be very different from another one. And I'm not sure as a clinician how to, sort of, know when I can trust those results, so to speak.

Mark Rubin: Well, I think, obviously as a clinician, you have expertise in things you're doing that hopefully there's a partner in this, that has expertise in interpreting the data, whether it's in pathology a molecular pathologist, for example. And so we think this is obviously a very important activity. At the AACR, for example, we're trying to support the role of pathology in this area, so I think it's got to be a partnership. We heard yesterday from Colin Pritchard, gave a wonderful presentation about MSI testing that you mentioned earlier. It's so complicated. I mean, I consider myself an expert in this area, and every time I'm always learning new things. I think we really have to be really consulting with the experts and we can't understand everything, and it's a dynamic process changing throughout.

Charles Ryan: Yeah. And finally, moving beyond just identification of a loss or a mutation, are we going to get to a point where we have functional assays? That's another one of these sort of holy grails, I think, of biomarkers, right?

Mark Rubin: Yeah. I mean getting back to Colin Pritchard's talk, I mean, thinking about DNA repair, we have to think about, are these mutations that we see functionally active or functionally relevant? Dr. De Bono brought that up yesterday as well. So this is obviously very important. There are groups that are developing assays to try and do that, and one can even imagine using patient-derived organoids, where you could potentially look at this, and this has been studied in colon cancer, so why not prostate cancer?

Charles Ryan: Yeah, I actually looked at the conversation about BRCA2, for example. In biallelic BRCA2 loss, you may have an exquisitely sensitive tumor to PARP inhibition, and that's great. But what about monoallelic loss? Is it simply less sensitive, apropos of your point earlier about making decisions based on AR status, for example? And those are really excellent questions that we need to answer in the clinic and with our pathology partners, basically looking at outcomes as a function of not just a binary mutation versus no mutation, but more qualitative analysis perhaps of those mutations.

Mark Rubin: Yeah.

Charles Ryan: All right. Well, thank you as always for your leadership in this area, and I'm looking forward to your talk and great chatting with you.

Mark Rubin: Same.

Charles Ryan: Yeah.