One of the most exciting areas is looking at the biology of cancer progression. There have been several important advances, most recently a paper in Cell a couple of years ago that looked at the molecular phenotype and genotype of lethal prostate cancer. This was a Stand Up To Cancer (S↑2C) project that brought some interesting and new insights about what really drives prostate cancer, especially in terms of the patients who are most at risk for progressing to lethal disease. The most interesting biological target was really DNA repair and pathway mutations, particularly BRCA2, which we had always assumed occurred in very small percentage of patients with prostate cancer. However, it turned out in this study that up to 25% of patients had either germline or somatic mutations in the DNA repair pathway. That’s a really large number; this is important not just because we can actually potentially test for it in the germline and in the tissue, but also because we might be able to target it with drugs like poly(ADP-ribose) polymerase (PARP) inhibitors or even platinum chemotherapy. I think that we’re going to see more and more use of molecular targets – and we’re certainly seeing it in other diseases like breast, colon, and lung cancers – and new drugs tested in subpopulations of prostate cancer rather than treating all patients the same with the same type of drug. The idea is that we would start to actually molecularly characterize them and treat only a subset of patients, for example, with a PARP inhibitor or carboplatin.

It really is intriguing, Dr Oh, that we’re starting to think about mCRPC in terms of the molecular definitions. I agree completely about the points you made about the S↑2C paper with the DNA repair defects. But the implications for some of these DNA repair defects, if they are germline or inherited, really extend to family members as well. I think that’s one of the newer recognitions. There was an article in The New England Journal of Medicine last summer from the Seattle Group and others that demonstrated a high prevalence, almost 12% of patients, who actually had germline alterations if they had metastatic disease. Not only does that have therapeutic implications, as you mentioned for things like platinum and PARP inhibitors (and there may be other ones, maybe even radium as well), but it also has real implications for the family members. Particularly if females end up inheriting a BRCA mutation and the penetrance is very high, there can be breast cancer, ovarian cancer, and even prophylactic mastectomy and oophorectomy, etc. But to kind of swing back to the segmentation of the castration-resistant prostate cancer (CRPC), researchers are also defining androgen receptor (AR) mutations that are very particular in how they will respond. There are things like a 702 mutation in the AR that tends to be activated by prednisone. There’s an 875 and 878 mutation that may lead to promiscuous receptors and interact with things like progesterone and estrogen in interesting ways. The implications are leaping out. Another one of interest was the phosphatase and tensin homolog (PTEN) deletion, which actually had an activation of protein kinase B (AKT) associated with it. Now there’s an AKT inhibitor that’s been in trial, and it looks interesting enough to continue the trials to go to the next step. There are lots of provocative things right now in terms of likely targeted therapy, and that doesn’t even mention the new therapies, like the prostate-specific membrane antigen (PSMA)-targeted therapies that we’re also starting to talk about. It really is an interesting and intriguing time right now.

I’m usually a very optimistic guy, but I just want to bring in a dose of reality on this though, if I may. Frankly, this is a field in which prostate cancer research is relatively behind. We’ve been doing a lot of molecular profiling in a lot of other cancer types for quite a long time, and we’ve developed a number of targeted therapies. I think one of the lessons we learn from fields like lung cancer and others is that it’s always more complex than it looks. So in the case of BRCA2, it may be more than just a point mutation that is needed. Maybe we need real loss of a whole gene, and maybe multiple hits are needed to get the deep, durable responses that have been reported. I very much look forward to the phase 3 data and the studies that absolutely have to be performed. In the field, we have to commit to doing these biopsies to find these patients and really show in a much broader setting than a few academic centers that these results hold true. But I just want to caution that this is not going to necessarily change the paradigm for the majority of our patients. It’ll continue to be a little bit of a zoo, and for the majority of patients, we are still going to need therapy that is focused on the whole disease burden in the population, like chemotherapies and radiation therapies. Here’s where I get excited, around what Dr Sartor has said about PSMA targeting, because that target is present in the vast majority of patients, even in this molecularly heterogeneous castration-resistant disease state. For immunotherapies, I am very excited about some of the data around androgen spreading and the opportunity for immunotherapy to begin to broaden its target across that heterogeneous disease burden. I think that those strategies potentially could result in a paradigm shift for a larger proportion of patients with prostate cancer.

I think we generally agree, but even in this era of molecular targeting – whether it’s with the first epidermal growth factor receptor inhibitors or any of these targeted therapies – we see that these therapies are only going to work in a subset of patients, even when the particular target may be widely expressed. Immunotherapy is a good example. Prostate cancer has not been one of the leaders in terms of checkpoint inhibitors, but even in the diseases where immune checkpoint inhibitors have had really dramatic responses, it’s only up to about one third of patients who have had those responses. If we double up on the checkpoint inhibitors, let’s say in melanoma, we can increase the response rate, but with greater toxicity and cost to the patient. In the end, I think we’re going to be stuck with subsets for a long time. I think this is true of PSMA targeting as well.

Let’s take AR targeting, which is probably the most important driver of prostate cancer. We have some very, very good AR-targeted drugs. But we also have always known that some patients get dramatic benefits from primary androgen-deprivation therapy (ADT), for instance, where they may not progress for a decade or longer, while others may progress within a few months, even though they both officially respond to that treatment. Without dividing these patients into molecular categories, we’re never going to really progress in the field. DNA repair pathways are only going to be active as drivers in a subset of patients, but we have to identify that population and start thinking of it almost like a different disease. It’s not prostate cancer; it’s like a BRCA-type cancer. That’s where I have more hope in that we are not just trying to target DNA repair pathways in prostate cancer. There has been lot of evidence in ovarian cancer and breast cancer suggesting that this may be a viable, clinically active, clinically relevant pathway in other patients. I think we have to be thinking broadly about the populations that we’re treating, in different categories across anatomic sites, whereby when we target certain drivers, whether it’s PTEN loss or BRCA mutations, that those drugs will work across those different areas. Immunotherapy really falls in that category because of that one third of patients who respond to, let’s say, the programmed death [PD]-1/PD-ligand 1 blockade. From the perspective of that target, they probably have more in common with lung and melanoma patients than they do with others from within the anatomic disease state.

I don’t disagree, and these aren’t mutually exclusive philosophies. I agree with the profiling – it’s just that it’s always going to be small subsets of patients. Regarding immunotherapy, there is a reason to be optimistic because there is now level 1 evidence from large, phase 3 trials with immunotherapy monotherapy, showing dramatic improvements in overall survival (OS) in an unselected population with kidney cancer. In bladder cancer, now we see the same thing, in an unselected population.

I think we disagree about the size of the subsets that may benefit from these drugs, but yes, improvements in OS are encouraging. Another consideration is that immunotherapies have, in lung cancer for instance, clearly led to negative clinical trial data in unselected populations. I think it’s a mixed bag with immunotherapy.

I think it’s absolutely true that we’re going to be able to have some targeted subsets, and it’s also absolutely true that many of these targeted subsets are not going to have effective therapies. But one of the interesting things that we’re finding about molecular profiling is that we end up with targets that are currently untargetable. For instance, how do we target a P53 mutation? How do we target the Wnt pathway in a successful way? We move down the molecular profile, and we find certain subsets where we might be able to actively and effectively intervene, also defining subsets where we really don’t have a lot that is in the current therapeutic pipeline. That’s where I’m very excited about some of these targeted radiopharmaceuticals. The interesting thing about radiation is that it rarely depends on any particular subset. Yes, we can have DNA repair and have a particular sensitivity to radiation in those DNA repair-mutated patients, but if we get the right dose of radiation to the right spots, we actually don’t depend so much on the underlying molecular defects. We can basically use radiation to pretty much kill what’s there. The problem is getting the radiation to the right spot. The PSMA, I think, is exciting because it’s often expressed in a number of these poorly differentiated patients who are responding poorly to things like enzalutamide. The fact is, enzalutamide may actually help regulate the PSMA. As we’re engaging in this repetitive and ever-expanding world of molecular profiling, I think we will have people who can benefit from the defects we find, and then people who are regularly going to require alternative therapies. To me, that’s really exciting because it gets back to the predictive biomarkers we were talking about a little bit earlier.