Oliver Sartor: Hi, I'm Dr. Oliver Sartor. I'm with UroToday here at the US Prostate Cancer Consensus Conference in Colorado Springs. Really a pleasure to be able to have Matthew Smith, professor of oncology at Harvard Medical School and at the Mass General Cancer Center. Welcome, Matthew.

Matthew Smith: It's good to be here.

Oliver Sartor: We're going to talk about an interesting topic today. We're going to talk about loss of tumor suppressor genes and the intensification of chemotherapy. Now, before we get into parts of the debate and what you might view as being a very interesting topic, let's just talk about tumor suppressor genes in prostate cancer then let's talk a little bit separately about intensification of chemotherapy because those are two important topics on this particular agenda.

Matthew Smith: Absolutely. So I think there's good research, I think best classified as biomarker discovery that has looked at loss of tumor suppressor genes as a prognostic marker, so predicting worse clinical outcome, and I do describe it as biomarker really development, because for the most part, those were not pre-specified analyses. They're not clinically validated. They haven't proven clinical utility. For example, it's not clear that compound tumor suppressor gene loss adds value in multivariate analysis to predict what we already know based on clinical criteria. But nonetheless, I think it's sort of commonly used for that purpose with a lot of caveats that should be considered. But perhaps the greatest unmet need is the one you alluded to, and that would be predictive biomarkers where we could use molecular information to identify patients who would be best treated in one fashion or another. And relevant to this question is whether they should receive chemotherapy at all, perhaps as part of initial therapy or intensification of chemotherapy treatment by addition of a platinum.

Oliver Sartor: Now, addition of platinum is intensification. Are there any other intensifications that you'll put into this bucket, or is it just platinum added to a taxane?

Matthew Smith: Well, that's why I think probably the more common question really is so-called triplet therapy, right? We have compelling evidence that treatment intensification for mCSPC improves progression for an overall survival. Most or all patients should receive ADT plus an ARPI. We now have evidence that triplet therapy is superior to ADT and docetaxel. We don't yet know. We don't actually have the converse data to say that the addition of a taxane improves survival in patients receiving a doublet of ADT and an ARPI. So we're left making decisions based on clinical judgment and perhaps is a very important area for biomarker development and validation. So if we knew who benefited most from triplet therapy, we'd be in a better position to make those recommendations. Now, national guidelines do have an opinion on this. So for example, NCCN guidelines recommend triplet therapy for patients with high volume, either synchronous or metachronous disease. So they do use clinical criteria to guide treatment choice, but there's not yet in NCCN guidelines any molecular predictive markers to make that specific decision.

Oliver Sartor: I want to drill down a little bit more in the tumor suppressor genes because we have a very broad audience, and I think I know what you mean and I know you know what you mean, but explain a little more detail. Which ones are you talking about? Which are the ones that are prevalent in prostate cancer that we need to be concerned about?

Matthew Smith: Yeah, so really there are really three tumor suppressor genes that we're talking about relevant to this conversation. So it's p53, RB, and PTEN.

Oliver Sartor: Okay.

Matthew Smith: And through biomarker discovery, research has suggested that loss of two or three of those tumor suppressor genes is associated with a worse clinical outcome than either having lost none or only one of those tumor suppressor genes. So by that biomarker definition, they're kind of weighed evenly. It is important to note, and I really liked your question because it really comes down to which biomarker are you talking about? Much of the work to evaluate prognosis has been next-generation sequencing of tumor, but there are other biomarkers that have been suggested that are either used different forms of testing, and I'll describe that in a moment. And they might be loosely lumped as tumor suppressor gene biomarkers, but for example, PTEN transcriptome analysis, very nice work done out of STAMPEDE, is not the same as the compound TSG loss biomarker that's been described using next-generation sequencing. And the available information suggests, for example, that there's considerable discordance between these biomarkers. I think really important to be precise about which biomarker we're talking about and how defined, particularly if it's being used for treatment selection.

Oliver Sartor: Right, it's very different than PARP inhibitors and BRCA2, so that's predictive. And I don't want words in your mouth, but a little bit earlier you said what we have are prognostic biomarkers, but we don't know if they're predictive or not. Is that an important distinction?

Matthew Smith: I think it's a really important distinction. Even their use as prognostic biomarkers, the clinical utility in that setting has not been established. So I think, look, we're all going to look at that information much in the same way we would look at patient age, overall health, anatomic sites of disease, disease volume, manner of presentation in making decisions about initial management mCSPC. But I think it's important to acknowledge that we need to avoid any suggestion that you simply look at next generation sequencing report and then make a decision about what treatment you apply because we are not yet at that level either for prognosis and certainly for prediction.

Oliver Sartor: Yeah, and earlier you alluded to the type biomarker. You mentioned DNA analysis, we can talk about RNA analysis, we can talk about protein analysis, we can talk about tumor analysis or circulating tumor DNA. I mean, all of these are a bit distinct and certainly they bear some relationship to one another, but they're not necessarily the same. I want you to elaborate, if you don't mind, on kind of the distinctions between the assays and the way we think about it.

Matthew Smith: Yeah, really super important point, and even for example, in NCCN guidelines as it relates to very nice work done at MD Anderson about the suggestion that tumor suppressor gene loss or a clinical phenotype of aggressive variant prostate cancer either or could inform chemotherapy intensification with cabazitaxel carboplatin instead of cabazitaxel. A couple of points here, NCCN doesn't specify which biomarker. The biomarker used in that study from MD Anderson was very specific. It was really biomarker discovery. In that study, they really only showed as a pre-specified analysis that the clinical phenotype of aggressive variant was associated with a better outcome by addition of carboplatin. All of the molecular biomarker work was in their own description, exploratory and iterative. They looked at IHC, they looked at CT DNA, and there was a lot of discordance between those two forms of assessment as well as with aggressive variant as clinically defined. The dilemma here is the following. If you accept any one of those criteria as reason to escalate with carboplatin, you would really almost exclude no one because of the considerable discordance between those definitions of aggressive variant.

Oliver Sartor: So if I'm sitting in the clinic and I'm a clinician and I'm going to say, "Oh, I've got a bad biomarker, I must intensify therapy," that may not be a good idea.

Matthew Smith: I think that's right. If it is put in the context of what else we know about the patient, including their manner of clinical presentation as we talked about in mCSPC, high volume low volume, de novo versus recurrent, and in the case of progressive disease, mCRPC aggressive variant, other distinctions, if it's put in that context, it may be a consideration, but it shouldn't be the sole driver. This is very different than drug selection, for example, with the DNA repair mutation treatment with the PARP inhibitor where we have very good information that that is a validated predictive biomarker, particularly BRCA2, for example.

Oliver Sartor: So, so far we've talked about the tumor suppressor genes, and the three you mentioned are the PTEN, the p53, and the RB, and then alone or in combination. And we talked a little bit about the assays, some of which may be tumor-based, some of which may be certain tumor DNA-based, some of which may be protein-based, some of which are genomic-based. So just kind of given the complexity for our listeners to be able to understand where we are. Now, your stance is a little bit of a debate that is not occurring because we're missing one of our other members, but that's okay, is that we should not be using these biomarkers to intensify. So if you will kind of encapsulate that argument and let me hear why you take that stance.

Matthew Smith: Yeah. So I think that at a high level, I do look at this information. I don't order genetic testing, next-generation sequencing for the sole purpose of looking for tumor suppressor gene loss. I'm typically doing it to look for validated predictive biomarkers, DNA repair deficiency, mismatch repair deficiency, where that information clearly can drive treatment decisions. But when you order those tests, you get other information, and that can be, I think, instructive when you put it in the entire picture of the patient to make, if you're on the fence about do you add carboplatin? Yes, no. Maybe that will be part of a reason to embark on that treatment course. Similarly, poor prognostic features by molecular testing might be a reason to consider addition of docetaxel as part of triplet therapy for mCSPC in a patient in whom you're on the fence.

What I'd like not to see is someone receiving that report and saying, this patient needs X because of that report alone. We're certainly not there yet. I think we need to do a lot more work to clinically validate some of these biomarkers, and it is very different. So this is not... We don't believe that these tumor suppressor gene products are the targets of the treatments that were describing.

Oliver Sartor: No, they're not.

Matthew Smith: They're not.

Oliver Sartor: And we know that.

Matthew Smith: Very different than the other examples I provided about DNA repair deficiency and mismatch repair deficiency. So totally different narrative. So bear that in mind. So I think we have a lot of work to do, and some of that work is ongoing, and I'm very glad to see that where there are attempts to prospectively evaluate these biomarkers and see if they really do make a difference regarding treatment outcomes.

Oliver Sartor: One of the things that worries me a little bit is when you have a new drug, and I'll say Olaparib, when it was new, there was an intensive desire to be able to validate those predictive biomarkers. The studies were done, large prospective trials with beautiful collection of centralized data, and you could really make conclusions and be able to make conclusions from a sponsored study with a significant new molecule is very different than taking an older drug like the carboplatin and then say, "Okay, we need to find those biomarkers." It's been a little bit disappointing to me that these drugs like carboplatin have been around for a long, long time, but we don't really have good predictive biomarkers for the use. That's been a disappointment for me. And I just wonder if you might think in part due of the lack of sponsorship, if we had a brand new shiny carboplatin that perhaps we might have more investment in the biomarker world.

Matthew Smith: It probably would be the case. And we didn't have those kinds of studies before they were without PARP inhibitors, right? You could have looked at carboplatin and DNA repair alterations long ago, but that wasn't done perhaps because I think there was a convergence of biologic understanding of the disease that led to that. And then there was the commercial opportunity to develop PARP inhibitors in prostate cancer. So I do share your concerns, but I think the good news is there are a number of studies that are ongoing to look at these questions. Typically, in the cooperative group setting, there's an Akron study called Aspire that's looking at whether compound TSG loss makes a difference for triplet therapy. So there is work ongoing. It's going to be a while before we have that result. So we're going to be left with many years of making clinical decisions without that high-level validation of the biomarker, if in fact those studies do validate them as useful biomarkers.

Oliver Sartor: If I were going to quibble a little bit with a question, I might say, why just intensification of chemotherapy and what about other alternatives that might also be considered? And obviously the PARP inhibitors are a separate class and they don't apply here, but we have radiopharmaceuticals who might be able to address these aggressive variants. And I felt a little bit pigeonholed by the question because I'm interested in maybe a larger menu of possibilities as we go forward. But that's just me, and maybe I'm not thinking right.

Matthew Smith: No, I think it's a great point. And I think, I suppose the question was posed this way because they're common clinical questions, particularly the triplet therapy question in mCSPC. This is something that you and I face every week in our clinic seeing such patients. And so there is some unmet need to make the most informed decisions, but there are certainly other value to carefully defining both prognostic and predictive biomarkers because you could identify basically poor risk populations who would be candidates for new targets or newer targets, novel uses or novel combinations of existing drugs. So absolutely, I think that's important.

Oliver Sartor: Matthew, we need to wrap up here in just a second, but I wonder if you might give a summary statement taking your view of the biomarkers, particularly the loss of the tumor suppressor genes and chemotherapy intensification, just kind of encapsulated parts.

Matthew Smith: Yeah, so I'll begin by just simply answering the question is, should compound TSG loss, meaning loss of two or three of RB, PTEN, or p53 change treatment recommendations? And I'd say no. The clear answer should be perhaps no, no, and no, because there's more than one biomarker we're referring to. But I think it can have a role as part of a consideration, but by itself should not be altering contemporary treatment decisions. We need much more work before we can make those kinds of recommendations. And so I'd be very careful in interpreting that information in clinic when you get it back with your next-gen sequencing.

Oliver Sartor: Great. Thank you, Matthew. Pleasure to have you on UroToday. Thank you for your insightful comments.

Matthew Smith: Good to be here. Thank you.