Christopher J. Logothetis, MD: These are exciting times for those of us involved in prostate cancer research.

Multiple new drugs have been approved, we face new challenges on how to turn these into therapies that alter the course of the illness, and even -- I think our panel here would agree -- we can now start having realistic dreams of increasing the cure fraction of patients with prostate cancer in a very meaningful way.

What we're going to address today are the challenges of both understanding these therapies and optimally integrating them into a rational treatment strategy.

First I'd like to build on the biological framework, which will help you logically understand how these therapies fit into treatment paradigm.

The basic assumption that governs what we're doing when we're dealing with multiple therapies is that the optimum integration of therapies will further improve the survival of patients with prostate cancer. There is a good way and a bad way to integrate these therapies. An effective therapy, given the complexities of what we have in front of us, has to be marker-driven, and these markers are going to be selected based on the understanding of the underlying biology.

As we move from a progression in a disease-state model that is driven by the practical and pragmatic issues of drug development, we want to move to a biology-based model that leads more efficiently to the combination of therapies.

These are the milestones in prostate cancer progression as we think about the interface between the biology of the disease and the clinical progression of the disease.

The primary tumor is often informative, and as many of us know, the findings and analysis of the primary are prognostic.

The behavior of the disease in between, as reflected in changes in PSA concentration, is predictive of the development of bone metastases or progression of the disease. Strategies at that point to inhibit progression to obvious metastases are being proposed.

Bone metastases are the hallmark of cancer progression in the advanced state. The importance of this is highlighted by the fact that when you target these metastatic sites, you can do something that cannot be done in other diseases -- that is, target the second echelon of metastases and prolong survival.

And then finally, you have castrate-resistant progression. We believe that there's still a role for androgen-independent progression, which is related to alterations in the cell cycle, and that complex microenvironment signaling pathways compose a network that drives the progression in an organ-specific fashion.

You could argue that optimum control of the primary site, as in all other diseases, has to be integrated into the effective treatment of metastases. I can tell you there are institutions, including ours, where the surgical removal of the primary, even in the presence of metastases, is being conducted in an exploratory fashion. And all the confusion and the controversy regarding how to integrate immunotherapy is a new challenge, because we're not yet armed with markers and the specific knowledge of picking the optimal interval. These are the issues we're going to have to address if we're going to make magic out of all the new therapy options that have been provided to us.

One of the biggest issues in the time-honored management of solid tumors is summarized in the following slide, and that is you optimize therapy in the advanced setting, like we do with bladder, breast, and all these other diseases, and then you progressively integrate it earlier in an integrated fashion with optimum control of the primary, and you expect an improved outcome.

But that paradigm makes one critical assumption that has not been tested in prostate cancer: that the events that occur in late disease are also reflected in the biology of the primary disease. So, if you're going to move up use of these drugs, you assume -- as has been true in other diseases -- that these relate. It turns out that in prostate cancer that may not be true. That's a critical distinction from some of the other common adult solid tumors.

On the left of this slide is a schema that we've all understood for many years, that 1 cell becomes 2, 2 become 4, 4 become 8, and they progressively increase in number and are progressively more heterogeneous. This increase in the number of heterogeneous cells is linked to a reduced responsiveness to chemotherapy. However, if that were true, then giving chemotherapy earlier, which works relatively well in the metastatic setting, should result in a further improvement in outcome. But the trial by Millikan and colleagues, which was reported 4 years ago in the Journal of Clinical Oncology and which had more than 300 patients, shows no evidence -- not even a hint -- of a benefit of chemotherapy combined with hormonal therapy as the primary treatment in unselected patients with metastatic disease. Furthermore, this is bolstered by the fact that neoadjuvant preoperative chemotherapy does not result in the expected reduction in antitumor activity.

This challenges the notion that the biology reflected in metastatic disease, which determines the response to therapy, is reflected in the primary disease. So the primary may be prognostic, but we have to examine critically whether understanding the primary tumor will also be predictive of therapies that we later need to give.

This slide shows the schema that we've proposed to summarize this concept. Basically what it shows is that, over time, increased chemotherapy efficacy parallels increased lethality. One would have expected that as a disease gets more heterogeneous with higher volume, the efficacy of chemotherapy would decline, but actually the relationship is the opposite.

We believe that there are data suggesting that pathways linked to progression drive through a phase of microenvironment dependence, which is relatively chemotherapy-resistant. This is analogous to how we think of chronic myelogenous leukemia, where early events do not allow you to predict how to treat somebody with blast crisis.

We hope that today's talks will not only give you a glimpse of the promise of new drugs that exist today, but a sense of the more complex therapeutic algorithm that we're going to need to attach markers to, to logically frame and apply these new therapies. This algorithm goes from a subset of patients who very early on need no intervention, to those who can be cured with surgical control of the primary, to the subset of patients who develop bone-forming metastases, and, finally, to these altered, more autocrine-driven altered cell cycle events that are chemotherapy-sensitive. This algorithm is not yet established, we don't have markers that support it, but it's increasingly supported at least by observations, which make this a logical stab as to what's going on with the disease.

You're lucky to have here today some of the people who have contributed greatly to the advances in prostate cancer: Dr Beer, from Oregon, who has expended a tremendous amount of time trying to understand and treat prostate cancer; Johann de Bono, who has done many of the most recent studies with some of the new novel agents that I know you regard as exciting and who comes to us from London; and Dr Fred Saad, who comes to us from Canada, who will speak to issues related to, and has championed some of the concepts of, targeting bone.

The first speaker I'd like to invite up is Dr Tom Beer, to discuss the current therapy in the integration of new therapeutic approaches in prostate cancer.

Tomasz M. Beer, MD: I really appreciate the chance to talk with you today. It's a wonderful time to talk about prostate cancer, and being with these guys on one stage is really one of the greatest honors that I can hope for in my career.

These are the things that I would like to touch on today. We're going to be talking a bit about targeting AR therapy. Johann is going to expand on that quite a bit, so we'll talk briefly about AR therapy, and, in particular, current emerging hormonal therapies. We'll talk about chemotherapy, immunologic therapy, an example of targeted therapies; we'll touch on bone-directed therapy, and try to sum it up as the emerging toolbox that we're dealing with.

There's a lot that can be said about the molecular justification or rationale for pursuing the AR, and we could give an entire talk on that.

But I just wanted to share with you one striking slide that shows that, in metastatic sites in castration-resistant prostate cancer patients, we can show all of the elements of an activated AR-signaling cascade, including the receptor, the downstream pharmacodynamic marker, and other androgen-responsive genes expressed. Other studies have shown, as Dr Logothetis has alluded to, both autocrine and paracrine androgen production. It's becoming increasingly clear that the androgen-signaling pathway is a key survival and growth factor throughout the lifecycle of prostate cancer.

Now, that's good news, because that's a target that we've understood a bit about for a long time, and that we can try to do something about now that we understand how important it is, even following the failure of frontline hormonal therapy.

There are a number of second-line hormonal therapies that have been in clinical use for quite some time. These include androgen receptor antagonists from the first generation, androgen synthesis inhibitors, arguably from the first generation, as well as the female hormones, the estrogens. All of these interventions share several clinical features. They produce modest but measurable and real tumor response rate, providing proof of principle that targeting the AR via these different methods can in fact provide antitumor activity following frontline castration therapy. None of these has also been shown to affect survival. To be fair, none of them have been studied in a way that would allow us to detect a survival advantage, but I think it would be fair to say that their level of activity is insufficient to make large phase 3 studies justified and likely to succeed.

I think recent years have brought us some real strides in the development of novel agents that exploit the signaling pathway, and that has really been an exciting development in the field. Two principal strategies focus on the ligand production side, and on better blockers of the androgen receptor. And those obviously are not the same; not all AR-mediated activity is ligand-dependent, and so I think that there's an opportunity here for different drugs targeting the same signaling pathway in a clinically meaningful, different way. Johann is going to show you all the detailed data, so I will show you just these summary slides regarding abiraterone and MDV3100.

The bottom line on abiraterone is that survival has been demonstrated to be improved in patients with metastatic CRPC following chemotherapy, and as a consequence, the drug has been approved and is available for use. And a second phase 3 study that is focusing on chemotherapy-naive metastatic CRPC patients is underway.

MDV3100 is in a similar situation. The initial results of this study were presented at this meeting earlier today and showed a significant survival advantage of 4.8 months at the median, about a 5-month progression-free survival, and a significant advantage with regard to PSA response rates and other measures of activity. The hazard ratio for survival was 0.63. These are very fresh results, and it is expected that these results will lead to an FDA filing and a second new-generation hormonal drug available to patients. Similar to abiraterone, there is a chemotherapy-naive study, and that study remains underway.

As all of the speakers will touch on, there are many questions, now that we have developed some better drugs. All of the basic questions that an oncologist would have asked with any new generation of drugs apply to these new drugs as well -- issues of sequencing, issues of combination, place of therapy, and the continuum of care. And I would speculate personally that these drugs could hold very substantial promise in the earlier stages of prostate cancer, where tumors tend to be maximally hormonally responsive, and I'm very interested in seeing what will happen when we essentially dose-escalate hormonal therapy, if you will, in the frontline management of prostate cancer, but we don't have any such data at this point.

And then of course there are mechanistic questions, I think it's worth remembering that there are paradigms in healthcare research that we can turn to and follow. When human immunodeficiency virus (HIV) drugs came along, we knew very little about mechanisms of resistance, drug selection based on mutational status, drug combinations, and so forth. And we need to take the same approach here, to better understand mechanisms of response and resistance, mix and match these drugs in a molecularly driven way, and hopefully yield the same sort of quantum leap in outcomes that we saw, for example, with HIV therapy. The challenge before us is considerably greater -- the HIV virus has a very small genome by comparison to human prostate cancer -- but the principle in the approach really should be quite similar.

Now I want to talk a bit about some of the other agents in use in this disease, and I'll mention each one of these quite briefly.

Currently there are 2 chemotherapy drugs that have been shown to improve survival in patients with prostate cancer.

I think there isn't a prostate cancer clinician alive who hasn't seen these curves: These are the original docetaxel curves. They show a modest improvement in survival over mitoxantrone. From the perspective of time they look modest, but it is important to remember that this was the very first randomized study to show that what was then called "hormone-refractory prostate cancer" can be treated with an improvement in survival. This is really what launched quite a bit of the work that has happened since, and it provides a historical context for where we are today.

Recently cabazitaxel, another taxane agent, was studied in the so-called TROPIC study. Many of you have seen these data as well. TROPIC was an international study of 755 patients who had been pretreated with docetaxel and were randomly assigned to either cabazitaxel or mitoxantrone, which at that point had been relegated to second-line status and was probably an appropriate control arm for this study.

This is the result of that study, a hazard ratio of 0.7. Given how short the overall survival in this group was, the difference at the median was relatively modest, under 3 months. These results led to the approval of cabazitaxel and I think it would be fair to say that for most of us, with regard to chemotherapy, docetaxel is the standard frontline agent today, and cabazitaxel is the standard second-line agent, as the only 2 agents to improve survival.

Now we're also in a new era insofar as immunotherapy is coming into its own in this disease, and there are many immunotherapy strategies being evaluated in clinical trials. I wanted to highlight 1 approved and 1 investigational method tonight.

Sipuleucel-T is a cellular adoptive immunotherapy. This is a schema of how it's carried out. This is a therapy that requires leukapheresis to collect mononuclear cells. That product is shipped at room temperature with live cells to a cell-processing facility, where it is incubated overnight with a prostate antigen tied to a granulocyte-macrophage colony-stimulating factor (GM-CSF) molecule, and then re-infused into the patient following this immune-stimulating procedure. The hope is that that this strategy will essentially teach your immune system, if you will, to recognize prostate cancer as something worthy of attack.

Sipuleucel-T was approved for use on the basis of this 512-patient, randomized study, carried out in asymptomatic or minimally symptomatic metastatic castration-reistant prostate cancer (CRPC) patients, with survival as the primary end point.

The survival was, in fact, better with therapy at the median, increasing by about 4.1 months from 21 to 25 months. Hazard ratio was 0.76.

We are not going to get into this right now, but a recent article in the Journal of the National Cancer Institute tried to make the case that the apparent positive results of this study might have been explained not by the beneficial effects of the active immunotherapy but potentially by the detrimental effects of the immune depletion carried out through the leukapheresis in the control group of patients who did not immediately get their mononuclear cells back.^[1] I'm going to take this highly controversial charged question and pick on some members of my panel at the end of my talk to address it.

Ipilimumab is a different type of drug. This is not a prostate cancer-specific drug. In fact, as many of you know, it was approved for melanoma when it was shown to significantly improve survival in that disease. This is a monoclonal antibody that targets the CTLA-4 antigen, a key cell surface molecule that mediates immune tolerance, one of those phenomena that makes sure that your immune system does not attack your own body. When one blocks CTLA-4, one can very potently activate the immune system.

What has been interesting to me is that, unlike some of the vaccination strategies, ipilimumab does produce real antitumor activity in prostate cancer. This is a waterfall plot of PSA changes that we saw in our phase 1/2 study that was presented by Susan Slovin from Memorial and that a number of sites around the country through the Prostate Cancer Clinical Trials Consortium participated in. So you can see that about a quarter of the patients had confirmed PSA responses, and any PSA declines were seen in about half of the participants across a range of doses of this drug.

The current status of this drug is that it is commercially available for melanoma patients. Sipuleucel-T is approved for minimally symptomatic or asymptomatic patients, and ipilimumab is approved for melanoma and phase 3 trials in prostate cancer are underway. You might detect a certain pattern here: there is a post-chemotherapy study and a pre-chemotherapy study.

Finally, I want to touch briefly on targeted therapy, and there are many targets that are being exploited with novel agents. None, perhaps, has gotten as much attention as XL184.

This is a combined Met and VEGFR2 inhibitor which was studied in this very complicated, randomized discontinuation design, which actually yielded interesting data prior to the discontinuation portion.

Many of you have seen these radiographs of bone scans with substantial improvements in the bone scan signal, or even resolution that occurred relatively rapidly -- as early as 12 weeks in some patients.

These are data on measurable disease, and what you see here is not too dissimilar from some of the kidney cancer signaling inhibitors. Not a large number of fully confirmed responses, but a large majority of patients with what appears to be tumor control, and some decreases in tumor mass.

If you look at effects on bone pain and narcotic use, these too are quite impressive. The pain improvements seen at week 6 and week 12 occurred in more than half the patients in this study. About half the patients who were on narcotics were able to significantly reduce their narcotic use. So along with these bone scan improvements, pain improvements were also seen.

One of the things that has been challenging with this agent is that the correlation between the various end points -- PSA changes and bone scans -- is less than perfect. The bone scan changes are so striking that they are almost unheard of. And for that reason this drug is not only exciting, but there is also a bit of controversy and uncertainty, and so I'm also going to pick on my fellow panelists to make some comments about how they view these results, and how they see this drug going forward.

The talk on current treatments wouldn't be complete without a mention of bone-directed therapy. As you know, prostate cancer is a bone-predominant disease, and there are 2 classes of agents that have been in development and that have either made it to the clinic or are likely to make it to the clinic in the near future: those that affect bone resorption, and radiopharmaceuticals.

Just 1 slide on skeletal-related event prevention, and phase 3 results of the 2 key agents in this space. Zoledronic acid was studied by one of my fellow panelists and showed a significant improvement in skeletal-related event outcomes when compared to placebo. It became a de facto standard of care. And recently denosumab showed an improvement over zoledronic acid. So, both of these agents have activity as measured by improvement in skeletal-related events.

The last agent that I would like to discuss is radium-223, which is a radiopharmaceutical that is yielding really unique results. This class of agents historically has been used for pain relief, but in this situation not only does it appear to offer pain relief, but also improve survival. These are data from a study of symptomatic, relatively heavily pretreated metastatic CRPC patients treated with radium-223 or with placebo, with an improvement in survival that is similar when you look at the hazard ratio to chemotherapy, and even approaching some of the newer AR signaling inhibitors.

Importantly, because this drug is bone-directed, it is gratifying to see that the effects are not just on survival but also on skeletal-related events, and here the hazard ratio is quite impressive, at 0.61.

The current status of radium-223, as you know from the slides that I've just shared with you, is that there is an overall survival advantage and the drug will be considered by the FDA for approval in the near future.

So the bottom line is that we are living in an era where there are more novel and promising agents under study than ever before. I've touched on some of them, and you're going to hear about more of them from my partners in this endeavor. This is really the first time ever that we've had 3 new drugs that improved survival, that are approved for use, in 18 months, and a couple of other agents already with survival data in the bank waiting for approval. So this really is an unprecedented time in that regard. We have the first-ever cancer-specific immunotherapy that targets prostate cancer providing us proof of principle that the immune system is an important target in this disease.

And I would argue that there are 2 major factors that have led to this progress. One is the rapidly accelerating understanding of the disease and its biology, and the other, which I think should not be underestimated, is the incredible collegiality and collaborative spirit of the genitourinary oncology investigators. I couldn't pick a tumor type to be in that has a friendlier or more collaborative set of friends and colleagues than prostate cancer.

I'm very interested in educating patients about clinical trials, and I wrote a book about that, which is coming up soon. If you are interested, our blog address is up there, come for a visit, and give me a holler if you have any questions about the book.

Dr Logothetis: So Johann, before we go forward with you, Tom brought up 2 questions. One relates to how do we deal with this critique that was raised about the immunotherapy, and the second one was, how do we grasp these striking, extraordinary bone scan responses with the Met inhibitor? Why don't you start with the second one first. Give me your view. Do you think there is proportionality here? Do they reflect the biology of the disease?

Johann de Bono, MD: The XL184 compound is generally having a major impact on patients' symptoms, PSA, and circulating tumor cell counts (CTCs). So the preliminary evidence in the phase 2 trials does suggest that the drug has genuine, bona fide activity. I think the bone scan changes are so different from what we are used to seeing with novel agents partly due to altered technetium uptake by osteoblasts. Therefore, the drug is impacting osteoblast function in some way, which may also be resulting in some of that symptomatic benefit. We'll have to wait on the phase 3 data, but I think that this drug generally has some antitumor activity.

We have published a study with the ARQ 197 c-Met inhibitor,^[2] and again have some evidence for falls in circulating tumor cells. We've also studied VEGFR inhibitors such as cediranib and again seen bona fide radiological responses. So I think targeting these pathways does have relevance in prostate cancer.

Dr Logothetis: One of the dilemmas that is emerging -- when you look at the recent report by Kevin Kelly with bevacizumab in prostate cancer^[3] -- is a significant prolongation of progression-free survival without a proportional impact on survival. And it raises the question: had we had a maintenance schedule to preserve that benefit, and had we thought of this like a hormonal therapy, would we have also gained a survival advantage?

Fred Saad, MD: I think that plays a role. But I think the other factor is what patients get after they leave. I think that, in part, can explain progression-free, and then if there are active agents our patients are able to be exposed to, you can, maybe, not reach the end point of overall survival. I think that is going to be the challenge with these agents that are brought earlier and earlier in the disease setting, with subsequent therapies that actually do make a difference.

Dr Beer: With bevacizumab specifically, I think one could make a very strong case that continued therapy would have been the preferred design given, if Kevin Kelly had the knowledge that is available today from other disease types. But there are other factors. You know, prostate cancer is a disease of a higher average age than most tumor types. Susceptibility to toxicity of these agents can also play an important role in the disconnect between PFS and OS. So in the case of bevacizumab, toxicity could have also been part of the equation.

Dr Logothetis: So this issue of trying to optimize a duration of therapy and adding on one layer of a therapy after another is likely going to become a challenge. We're going to have to figure out some strategy to limit or prioritize combinations of drugs, either within a specific set of progression of the disease, I would argue, or between patients, for the very reasons that you outlined. But I would think it's the consensus of the group that this issue of whether we maintain a therapy in the absence of continued evidence of regression has to be addressed.

Let me frame a controversy that some of you may not be aware of. Sipuleucel-T in a randomized trial worked to prolong survival. Many of you have struggled, as I have struggled in the clinic, to try to explain to a patient why he is better if the drug works and he is better if the drug doesn't work in any visible way. It is one of these therapies where you give it, and you run away and you declare a success and you don't know how to link a benefit to it. It has raised a huge challenge. And, in my mind, this theory of immunodepletion as being a potential cause for the improved survival by reducing the survival of the control arm is actually a manifestation of the fact that we have difficulty demonstrating evidence that an individual patient benefits. But I would argue, in talking with our immunotherapists and immunobiologists, that it would seem inconceivable, although it cannot be excluded with absolute certainty, that removing lymphocytes and exposing them and not returning them would result in a meaningful immunodepletion. It's such a small portion of the active cells that most immunologists discount it. But I think this controversy is raised as a symptom of the inability to link individual patient benefit to this outcome, and I've seen it in the clinic. I don't know if any of you feel that way.

Dr Beer: I think the other issue is frankly that there haven't been a lot of additional data forthcoming. I think that in the ideal world, where there aren't resource constraints, if a drug comes out that looks quite helpful but there are questions to be answered, then we go out and answer those questions. And then either find out that the drug isn't as useful as we thought, or we get reassured about the mechanism of activity or patient selection. But with this drug, since its approval we have basically had no new knowledge, these questions persist. And I think we desperately need new knowledge to tackle our understanding of the timing of the immune response, relationship to steroids, subsequently patient selection, refinement of mechanism of activity -- all of those things really ought to be answered with good studies.

Dr Logothetis: So the skepticism that has emerged around this, if I'm framing Tom's concerns right, is driven by the absence of data, which reinforces the need for randomized trials that have given us confidence to sustain this effort -- but it generates this uncomfortable feeling, when you're giving this drug for which you can't demonstrate a benefit.

I'm going to move on with our next speaker.

Dr de Bono will speak about the exciting events that he has played a central role in: modulating androgen signaling in the context of castration-resistant prostate cancer.

Dr de Bono: Thank you, Chris. Good evening.

We are making significant progress for treating our patients suffering from prostate cancer.

I want to talk to you about what's really been unprecedented. When I moved from San Antonio, Texas, working not far away from Chris in Houston, to London, England, about 8 years ago, I really couldn't have dreamed that in the space of 8 years we would go from 0 to 6 drugs that improve overall survival for this disease, when before that many phase 3 studies had failed to show a survival benefit.

Certainly, for hormone therapy, these advances have been fueled by a much better understanding of the biology of disease. First of all, we now have, I think, incontrovertible evidence that this disease is not hormone-refractory or androgen-independent disease. Moreover, I think we have incontrovertible evidence that bicalutamide and LHRH analogues do not provide maximal androgen blockade. I think we've also rewritten the true meaning of androgen deprivation and maybe even the meaning of castration in a patient with regards to hormone deprivation. Importantly, it all starts with the biology.

One of the key areas of progress, I think, has been this elucidation that androgens probably are key to causing prostate carcinogenesis. To simplify that, what we have is that androgens bind the androgen receptor, induce transcription in nuclear transcriptional islands, and through that process, perhaps in the presence of DNA repair defects, result in DNA damage that can sometimes be poorly repaired, leading to DNA rearrangements. What you end up with are rearrangements that make potent household genes to become oncogenes -- oncogenes that are driven by hormones. So the hormone is the foot on the accelerator driving the oncogene like ERG, ETV1, B-Raf, to drive and power tumor cell to grow. We know there are at least 20 different types of prostate cancer, many of which, maybe even most of which, are hormone-driven in this way.

We accumulated evidence over 20, 30 years that most mechanisms of resistance increase this signaling steroid-like androgen receptor axis, with increased expression of AR-activating mutations, DNA amplification resulting in the increased expression, and, also, importantly, intratumoral production of hormones by cancer cells, with evidence accumulating that a cell can make testosterone from acetate.

Within weeks of castration therapy in a patient, there can be sustained high levels of hormones inside that prostate cancer, despite ablation of testicular function. And maybe that's why many of our patients still have undetectable hormone levels despite LHRH analogues or surgical castration. And certainly these hormone levels, we believe, are still driving those rearranged hijacked oncogenes in these rearrangements. A bit like Bcr-Abl in chronic myeloid leukemia that we think is driving tumor growth.

So I'm going to focus on 2 drugs, one of which is already approved. The other we hope will get approved in 2012.

Many of you may be already using abiraterone acetate. Abiraterone acetate is a drug made by chemists at my institution, the Institute of Cancer Research and the Royal Marsden in London. And essentially this drug blocks hormone synthesis in the testes, in the adrenal glands, as well as in the tumor itself. And in fact, Charles Huggins, back in the '40s and '50s, postulated that indeed adrenalectomy would actually benefit patients with advanced prostate cancer. And I think our data have now in many ways confirmed his hypothesis at that time.

We tested abiraterone and published several papers with it in JCO and other journals, showing that this drug has substantial activity both in the chemotherapy-naive and post-chemotherapy setting. This may sound, in retrospect, fairly obvious now, but actually it was quite difficult to convince the community back in 2004, 2005, that it was worth studying a hormonal agent after chemotherapy. But can I state again that your decision and my decision to administer chemotherapy does not necessarily change the biology of advanced prostate cancer. If that cell requires those steroid ligands, that nuclear steroid receptor for its survival, your decision to give docetaxel does not change that cell's hormone dependency.

This is the compound; it was actually generated as a derivative of pregnenolone, a steroid derivative of dihydrotestosterone by our chemists. It's a potent and irreversible inhibitor of an enzyme called CYP17.

For over a decade, abiraterone lay undeveloped because many were concerned this drug would result in fatal adrenal insufficiency. I should remind you, if you remember from your physiology studies in medical school, that children born with an autosomal-recessive deficiency of this enzyme do not get adrenal insufficiency at any time. In fact, what they develop after puberty is impaired sexual differentiation as well as -- in some cases, but not all -- hypertension and hyperkalemia due to the high ACTH-driving mineralocorticoid synthesis.

Now, in our first phase 1 study, we initially deliberately added a mineralocorticoid antagonist, eplerenone -- not steroids -- to prevent toxicity associated with that mineralocorticoid excess, because we wanted to convince the community that this drug, without steroids, had antitumor activity. However, we now are administering this drug with steroids in studies, but I would beg you, do not give abiraterone without either steroids or eplerenone. I would also beg you to avoid using spironolactone, because this is a potent activator of the androgen receptor. Do not use spironolactone in patients with advanced prostate cancer.

We showed that abiraterone had antitumor activity, was safe, it hit the target, hormone levels fell, BHT became "super-castrative," if I can call it that, and the other hormones below the enzyme blockade fell, and hormones above the enzyme blockade increased.

And the rest is history, I guess: impressive PSA falls...

impressive radiological responses...

and impressive bone scan regressions.

I think for me the most gratifying results were seeing patients' pain go away and seeing opiates discontinued. One of my first patients on the phase 1 trial went from being really in severe pain on opiates to being pain-free and going to see his daughter in Australia, and that was just very gratifying.

There were lymph node regressions in many patients, and, in some patients, even 99% falls in PSA despite multiple prior lines of hormone therapy in this trial, up to 4 prior lines of hormone therapy before they got this drug.

We saw activity in the postchemotherapy setting with similar gratifying PSA falls in some patients by 90% in the post-docetaxel setting.

This led to this phase 3 study that we published in the New England Journal of Medicine. Several of the co-authors are here, my dear colleagues; we did this work together internationally. The randomization was abiraterone to placebo, in patients getting prednisone, 5 mg twice a day, and it was a 2:1 randomization, the largest and fastest randomized phase 3 study in advanced prostate cancer.

Interim analysis showed a significant survival benefit and this was confirmed at the final analysis, with a 4.6 month survival benefit.

And it did not matter if you had 1 or 2 prior lines of chemotherapy, it didn't matter whether you had pain or no pain, performance status, etc. The benefit really was present across all patient populations, young and old, and so on.

We had similar superiority for the treatment arm, for all the secondary end points.

Importantly, the drug was well tolerated. In fact I can honestly say that I genuinely could not distinguish whether a patient was on the abiraterone arm or the placebo arm. Similarly with MDV3100, I couldn't really guess. In fact, quite often when the trial was unblinded, I realized I had guessed wrong which arm the patient was on, because there was minimal easily distinguishable toxicity.

There are some adverse events of special interest associated with both of these drugs. With abiraterone it can be fluid retention, hypokalemia, uncommonly transaminase elevation. But I should say if you look at the grade 3/4 toxicity incidence frequency, it is very low.

Importantly, abiraterone does improve quality of life. It decreases pain, decreases fatigue, and these are the things our patients want to hear from us when we are recommending these therapies.

What about MDV3100? I was so excited to see the data become public today. You know, we are making major progress in this disease. And I should say with abiraterone we now have this drug in an adjuvant study in Europe, the STAMPEDE trial that is rapidly enrolling 300 patients a month. It is my hope and dream that maybe, in 3 or 4 years' time, we can say that we can improve outcome for high-risk, locally advanced disease with abiraterone, maybe even MDV3100. As you said earlier, Tom, that is where maybe we'll make a major impact, as trastuzumab has done in earlier stage HER2+ breast cancer.

MDV3100 is a novel oral androgen receptor signaling inhibitor, very different from bicalutamide, flutamide, or nilutamide. This drug has a profound impact on AR signaling at multiple levels.

The initial phase 1/2 study showed spectacular antitumor activity, very similar to abiraterone with regards to radiological responses, PSA responses, and circulating tumor cell count falls, with these associating with survival benefit.

And these are some of the initial data, about 50% to 60% response rates by PSA, or I should say PSA declined by 50%, with an impressive median time to radiographic progression, as you can see in the slide.

You know, just as an anecdote, I have patients at my centers doing abiraterone after 6 years. I mean, I saw a patient last week who's on the drug for 6 years and still going on drug. Now that is unusual, I have to admit, but these drugs generally have major benefit for our patients. We need to identify which patients particularly benefit.

This MDV3100 phase 1/2 study led to a randomized phase 3 trial that was presented today at GU ASCO by my colleague and co-chief investigator Howard Scher.

It had a similar, almost identical design to the abiraterone phase 3 trial, another trial with more than 1000 patients. This is the randomization, 2:1.

I don't have time to go through this trial in great detail, but essentially these are the primary and secondary end points. Overall survival is the key, the gold standard end point for a phase 3 trial.

At the interim planned analysis, the IDMC recommended unblinding of the study and that patients on placebo should be crossed over to the MDV3100 arm.

And this is another beautiful survival curve for a CRPC phase 3 study with a hazard ratio, as you've heard already, of 0.63 and a 37% reduction in risk of death. Maybe I would be over-claiming the data here, but I think we have some evidence here that patients are living longer than ever before. And certainly we hope to present at ASCO^® our data from the last 500 patients we've treated in our clinical trials program at the Royal Marsden. Our evidence suggests that patients with CRPC are living longer than ever before.

So future challenges -- what next? I think that the revolution in next generation sequencing and understanding disease biology is really going to impact our further armamentarium. I would envision that PI3K/AKT inhibitors will get approval for this disease. Combinations of MDV3100 and abiraterone, maybe ARN-509 and abiraterone, are going to be impactful in this disease. We're hopeful that the cabozantinib (XL184) data will bear fruit.

But, essentially, this is going to raise challenges for you and me: how do we treat our patients? How do we sequence our drugs? Your patients are going to say: what do I get next, doctor, cabazitaxel, MDV3100, or abiraterone? I think in 2012, it is possible that abiraterone will go into the prechemotherapy space, based on the COU-AA-302 data. But, most of all, I'm hopeful that we'll get abiraterone and MDV3100 into the adjuvant setting to really perhaps improve cure rates in high-risk, locally advanced disease.

So in conclusion, advanced prostate cancer is neither hormone-refractory nor androgen-independent, and commonly remains driven by nuclear steroid receptors, even, probably, after abiraterone and MDV3100. Hormone therapy works after chemotherapy, and our old-fashioned term maximum androgen blockade was and is a misnomer. Thank you for your attention.

Dr Logothetis: So Johann, can you speak to some issues before we go forward on prospects for combining anti-androgens or androgen-inhibiting strategies? Do you have any ideas of when and what combination you would prefer, if you were allowed to do anything you wanted?

Dr de Bono: The key issue here is that drugs like MDV3100 and the ARN-509 compound which we are excited about -- they're very similar drugs, they actually vary only in 1 residue. These drugs are competing with dihydrotestosterone to block AR, and unfortunately dihydrotestosterone is 10-fold more avid for AR than these compounds. So we hypothesize that if you add a drug like abiraterone and remove those hormones -- DHT, testosterone -- you will substantially improve the potency of drugs like MDV3100 by decreasing the need of the MDV3100 to compete with those nuclear steroid ligands. So I would envision that a trial of abiraterone and MDV3100, versus abiraterone alone or MDV3100 alone, would be substantially superior for overall survival benefit. And we have preclinical data to support that.

Dr Logothetis: I can tell you that trial is actually starting at MD Anderson next month.

Next, Dr Saad will speak to us about issues regarding the signal events that drive cancer progression in bone, with a specific focus on sort of the "poster child" for the 2-compartment model, Src-signaling. Fred?

Dr Saad: Thank you, Chris, and thanks for the honor of being with this distinguished panel.

We're going to backtrack a little bit and go back to some of the preclinical and early clinical validations of targeting, what we think is, an extremely interesting and potentially paradigm-shifting approach to prostate cancer. We know that we have bone-directed therapy that can control a progression in terms of symptoms and SREs. We have drugs that control the cancer and help to prolong survival. I guess the ideal would be to target both compartments -- both the proliferation of the cancer and also the metastasis.

If we take the model of Src and Src-family kinases, they play a key role in cancer signaling pathways through different mechanisms.

To summarize, Src-family kinases play a role in invasion, tumor spread, chemomodulation, and malignant bone disease, which are all important for prostate cancer control, both of the tumor itself and the preferential site of metastasis.

There is evidence that Src-family kinases play an important role in the transition from androgen-sensitive to castration-resistant disease in the preclinical setting.

And there is rationale for inhibiting the Src and Src-family kinases because of their activity that's increased in prostate cancer, and the role they play in regulating osteoclast and osteoblast function. Thus, adequate inhibition of Src and Src-family kinases should lead to tumor growth reductions, reductions in tumor metastases, reductions in areas of bone absorption and related complications, as well as tumor angiogenesis.

There are several Src and Src-family kinases inhibitors that are being investigated at different levels: bosutinib in preclinical; KX2-391 in phase 2; saracatinib in phase 1 and 2 as monotherapy, and starting in combinations; and dasatinib, which really targets the most of any of these inhibitors, has completed phase 1 and phase 2 as monotherapy, and also in combination. This is the most advanced of the different inhibitors, having completed a phase 3 study that I'll briefly touch on at the end.

Bosutinib blocks prostate cancer growth in bone metastases in the preclinical model, shown in vitro and in vivo.

The KX2-391 substrate-binding pocket inhibitor has a wider spectrum of in vitro activity in solid tumors and is more potent in the mouse xenograft models than other Src/Abl kinase inhibitors. It also inhibits microtubule polymerization. Phase 1 data has demonstrated biological activity at least in 1 patient with significant PSA reductions. The DLTs are reversible within 7 days based on the available evidence, and the phase 2 study is just completed with data still pending.

Saracatinib inhibits PC-3 growth and invasion in the preclinical setting. This was elegantly demonstrated in the work by Yang and colleagues.

In the in vivo model there is evidence of reductions in osteolysis in the bone, both radiologically and in terms of markers of bone resorption such as serum CTx and urinary NTx, which are reduced in a dose-dependent fashion.

In the phase 2 setting, there is a completed study where 5 of 28 patients had a PSA decline. However no patient achieved, unfortunately, a 30% decline. The median PFS was about 8 weeks, and there is at least some rationale, in the preclinical and in the clinical setting, to go forward maybe in earlier stages or consider combination therapy. As far as I know, there is little progress with saracatinib at this stage.

Dasatinib, with what we know in terms of the effects on Lyn and Src, potentially may have tremendous impact on reducing tumor invasion and tumor spread, modulation of the chemotherapy, and effects on malignant bone disease through Src-dependent bone resorption.

In the preclinical setting, there is clear evidence that it inhibits prostate cancer cell adhesion and invasion.

It seems to have a significant effect on proliferation and migration.

Some elegant work coming out of Chris's lab has shown a dose-dependent reduction in the number of metastatic lymph nodes and the size of the metastatic lymph nodes. Thus, there is antitumor effect over and above the effects in the bone.

There is clearly an osteoclastic effect, and experiments have shown that it is clearly more osteoclastic than osteoblastic, especially when compared to other agents such as zoledronic acid.

And it blocks osteoclastic formation and activity in a dose-dependent fashion.

There is evidence in the human model, based on our exploratory analysis of the effects of controlling urinary NTx, one of the best resorption markers, where reductions or normalizations of these resorption markers lead to reduced bone complications, but also to improved survival, in a purely bone-targeted approach.

So what is the clinical evidence for dasatinib as monotherapy in CRPC? There are completed studies in which 25% of patients responded or had stable disease when used as monotherapy. PSA doubling time improved in 88% of evaluable patients, with 1 patient having a clear PSA response. In terms of the bone, 51% of patients experienced a greater than 40% decrease of urinary NTx, and this was the same regardless of whether they got bisphosphonates. And 59% had experienced a decrease in BAP levels, again, with or without bisphosphonates.

In terms of waterfall plot, the decreases in PSA are illustrated here, and again, this is in the monotherapy setting.

Decrease in urinary NTx was even more impressive, again, regardless of whether bisphosphonates were used.

In terms of tolerability in the monotherapy setting, given as a daily dose, there were relatively few grade 3/4 events. However, in grade 1/2, there are cases of pleural effusion, dyspnea, and some GI toxicities, but very few grade 3/4.

So this led to the concept of combining dasatinib with docetaxel, with the rationale that I introduced the talk with. And there is evidence that there is synergistic effect of this combination, compared with using dasatinib alone, or docetaxel alone.

There is a phase 1/2 study of the combination that was done in 16 patients with 5 dose cohorts, and an additional 30 patients with 100 mg per day dosing with docetaxel every 3 weeks.

And in this study, 80% of patients had a decrease in PSA from baseline, with 57% having a confirmed PSA response sustained for over 6 weeks.

So here we see quite impressive PSA declines, obviously without a control arm, but more than what we would expect with docetaxel alone.

In terms of RECIST criteria, it becomes quite interesting. There were 30 patients in the study who had measurable disease, and, of those, clearly 60% had a PR, and 17% had stable disease, with an overall disease control rate of 77%.

This recently published study showed that, in terms of the waterfall plot, there were significant reductions in tumor size in patients who got the combination of dasatinib plus docetaxel.

Urinary NTx was very well controlled by this combination, unlike when docetaxel is used alone.

Here we see also an interesting reduction in BAP levels, compared to baseline; 75% of patients actually experienced a reduction in BAP. Alkaline phosphate levels correlate almost directly with BAP. Many of the studies highlight the importance of alkaline phosphatase control in patients on chemotherapy, and elevated alkaline phosphatase is a very poor prognostic marker for patients with metastatic CRPC.

To illustrate some clinical examples from the phase 2 study that were shared from Chris's institution, with dasatinib plus docetaxel, there were bone scan improvements and stabilizations that lasted more than 2 years.

When looking at PSA responses with docetaxel, there were significant reductions, and patients who were maintained on dasatinib continued to maintain significant reductions and stabilization of their PSA.

In terms of measurable disease, clearly in this patient there were significant reductions or almost disappearance of metastatic disease in the retroperitoneum. And again, the PSA, even after stopping the docetaxel and maintaining the dasatinib, maintained long-term control and stabilization -- coming back to the maintenance approach to these kinds of therapies.

In summary, dasatinib and docetaxel combination treatment was well tolerated. Responses demonstrate the antitumor activity of dasatinib in combination with docetaxel with 30% reductions in size and/or number of lesions, and 60% having confirmed partial response by RECIST criteria. Dasatinib favorably modulates bone turnover as a single agent and in combination, with 76% of patients having decreases in BAP from baseline, and 47% having greater than 35% reductions of urinary NTx compared to baseline.

This led to a phase 3 study, the READY trial, looking at almost 1400 patients who were randomized to docetaxel plus prednisone plus placebo, in the standard Q3 regimen, compared to docetaxel plus prednisone plus dasatinib at 100 mg daily, with a primary end point of overall survival. Secondary end points looked at changes in bone markers, time to first SRE, changes in pain intensity, and time to progression, tumor response rates, and safety and tolerability.

In conclusion, targeting Src and Src-family kinases in prostate cancer has a critical role in prostate cancer, due to the effects on tumor proliferation and metastatic bone disease. Dasatinib is a potent inhibitor of multiple Src-family kinases, with direct antitumor activity, which is critically important; synergy with tubulin-binding drugs; and direct effects on osteoclast inhibition. There is definitely phase 2 evidence of activity in metastatic CRPC. The randomized phase 3 trial is completed, and the results are pending. Thank you very much for your attention.

Dr Logothetis: So what I'm going to try to do before we go on with the other issues is try to integrate these 3 talks into some cohesive approach, and then see if we can answer some questions.

If we go back and we think of prostate cancer, and we think of changes in AR signaling over time, and we reclassify it based on states of disease, from the perspective of androgen signaling, this is how you might think of it. In very early disease, there is dihydrotestosterone-dependent cancer. What is the evidence to support that?

The evidence is from the PCPT trial^[4] and the recent bicalutamide trial,^[5] which basically showed that low-Gleason score tumors, but not high-Gleason score tumors, were avoided by the use of finasteride in preoperative treatment. As we already know, even in the presence of low-grade disease, at present, finasteride does suppress low-grade disease and has no effect on high-grade disease, or may even promote its progression, as some would suspect.

This indicates that there is heterogeneity in response to DHT within tissue, and there's a subset of these morphologic cancers that are DHT dependent.

Now the next step is the one that Johann spoke of in this new perception, and that is that the transition from endocrine-regulated prostate cancer to paracrine-regulated prostate cancer has replaced our old theory of castration. This is the new milestone that transitions from a cancer that is nonindolent to one with lethal potential. So a critical milestone for the progression of prostate cancer to one with lethal progression would be that it's able to respond with a paracrine-mediated progression, or endocrine-mediated progression, under the selective pressure of castration.

So this leads to the second category that reflects this transition, and that is the androgen signaling category.

What is the evidence that there is a subset of patients who are dependent on these pathways? It's reflected in the historical data that approximately 20% of patients with traditional castration do well with control of their disease over the sustained period of time.

Here is an example of a patient who evolved from intermittent hormonal therapy. He had a rise in his PSA concentration late, despite castrate levels of testosterone, but bone marrow testosterone levels started rising. Abiraterone acetate was introduced, and this man has now almost 20 months of undetectable PSA, even at a super-sensitive level detected and was treated right as he transitioned from endocrine- to paracrine-mediated progression.

So the third category that broadly describes prostate cancer progression is characterized by this broader network -- where Src plays an important role -- of interactions between AR and non-AR pathways that requires combination therapies. This is the microenvironment-dependent subset of patients.

Now we believe that Src-family kinases are the poster child of this. Phospho-Lyn is initially present in the stroma in the primary tumor under the normal state, but not in the epithelial compartment. Phospho-Src and phospho-Lyn are co-expressed in widespread castration-resistant disease.

This suggests, and this profile is consistent with the profile of a 2-compartment model, that there is a stromal event and a complementary epithelial event that drives the progression of the disease. So you could call this sort of a Knudson's hypothesis at a cellular level, where, in the primary and normal condition, neither the tumor nor stroma are able to exhibit an interaction, and then one or the other exhibit this pathway that's altered, making it sensitive for a combined effect, resulting in an organ-specific progression of the disease. And it's perfectly illustrated by the Src-family kinases in progression, and accounts for the need to both combine chemotherapy with dasatinib, because you need an epithelial-targeting agent, when you develop these therapies. The concept here is that you have to co-target both the epithelial and stromal compartment.

Now, just as we thought we were smart and we thought we were going to peel back this disease by narrowly focusing on microenvironment events, we realized that things were becoming different.

First of all, there's this obvious evidence that when you transition a tumor from a human background to a murine background, as you develop the xenograft, as many of us are, one of the most immediate things that happens is that the xenograft tumors are underrepresented in androgen signaling. This indicates that this tumor, under the selective pressure of this transition, is able to survive in the absence of androgen signaling in a significant proportion.

So, the more successful we get -- unless we treat them early -- we will again likely have to revert to the concept of androgen-independent disease as a challenge that's going to haunt us in the future.

So we've actually pursued this with some vigor at our institution anticipating that this is the next problem, and we've developed a clinical trial that is run by Dr Ana Aparicio and a more basic initiative by Dr Sankar Maity. They showed that during a castration-resistant progression, specific neural genes are elevated and there's alteration in cell cycle. And these events predict for true androgen independence.

And again, the loss of RB, an absence of AR signaling, and alterations of cell cycle are the hallmarks of this aggressive disease that emerges late in most patients and comes up early in some patients, and is not always correlated with small cell appearance under the microscope. It can clinically look like small cell, without being bone-homing, but when you biopsy it you constantly see morphologically an adenocarcinoma.

There is a manuscript that recently has been published showing that there is activation of mitotic programmed kinases that characterizes this, and identifies new therapy targets, and polo kinases and aurora kinases are actually being targeted for this subset of patients early on in combination with chemotherapy.

Such trials are being activated in our clinic, targeting these subsets of patients whose disease, we believe, is driven by altered cell cycle and who are truly reflecting an androgen-independent phenotype.

The other thing that one would expect is that they would have a high proliferation and be responsive to cisplatin, or carboplatin, an agent that is not typically considered as approved for prostate cancer. And, in fact, that's exactly what you see. If you treat patients with the clinical phenotype associated with this molecular profile, and you give them platinum-based therapy, this is the difference in the survival, over the expected survival, if you use a TAX327 with Halabi nomogram.

So there is an increased responsiveness to this, there's a higher proliferation rate with this tumor, and they do not appear to benefit from AR signaling inhibition.

So, the overall schema that is linked to the milestones of prostate cancer progression -- trying to de-convolute prostate cancer through the prism of androgen signaling -- is as follows: we evolve from very simple androgen signaling pathways that drive the disease, DHT-dependence, to androgen dependence (and we know castration is superior to DHT inhibition with 5-ARI in established disease) to more complex microenvironment-dependent event, (which is an interaction of a broader network of AR-associated signaling pathways, of which one is Src-signaling and we're looking at it in combination therapies), finally to an epigenetically suppressed, or altered progression, to alterations in the cell cycle that drives this disease and needs markers to differentiate therapy.

All of us have been looking at prostate cancer early to try to predict what happens late. We've been emboldened, because you can prognosticate on your primary. But I'd like to point out that prognosticating, predicting if somebody's going to do poorly, is entirely different than predicting what they'll respond to. And the prediction of response from the primary tumor has to be based on the assumption that the biology in the primary reflects the biology in metastatic sites.

If this model of temporal heterogeneity is true, if therapeutically meaningful differences in the biology of the diseases are emerging over time -- which we believe account for the responsiveness of chemotherapy late versus early, and if only tissue collected approximate to the intervention can be predictive and is informative -- then the whole idea of using the primary site to predict what will happen to metastases may be flawed, and we have to reexamine it.

There's been great skepticism on doing bone biopsies on patients with prostate cancer. But anybody who thinks it's not painful to do 12 biopsies on your prostate with a trans-rectal ultrasound -- and who then claims "But it's irrational to do a bone biopsy" -- lives on a different planet than I live on. We do bone biopsies daily on these patients. The business end of this disease is in bone, it's like leukemia; that's where you need to look. It makes no sense, in my mind, to try to look at the primary site and predict metastases and to justify not biopsying the primary, because somehow it's "more morbid." Just go into the ultrasound suite and see what happens with biopsies there, and then tell me if you wouldn't prefer a bone marrow biopsy.

So, we've got therapy targets in the "land of plenty" that are really causing us a huge problem. But this is the kind of problem that we all dreamt about, as Johann said. Not long ago, nobody could have believed it. We have novel microtubule poisons, but we don't know how to prefer one over the other and pick one at the right time with patients. We have alkylating agents that have not entered into the clinic and are not widely reported, because in the average patient with prostate cancer they have very modest activity; but most of us have come to believe that there is a subset of patients who benefit from these. There are bone-homing radiopharmaceuticals; there is immunotherapy; the checkpoint blockade and vaccines; the stromal-epithelial interacting pathways that entered into the clinic. Src-family kinases are being targeted with dasatinib; hedgehog; FGF; Met; VEGFR2. All these are in the clinic, and all these in phase 2 are demonstrating promising antitumor activity, and they're only going to work in the subset of patients.

And then we've got the signaling networks that we're also looking at in different combinations. Many of us are trying to block GSK-3-beta as a common node downstream with AR, and then optimizing AR signaling.

But unless we develop a marker-driven strategy to apply these therapies with some reasoning, I think the promise of these drugs is not going to be realized. So the plea is to do the kind of informative studies that look toward rational combinations that we have to build.

The model of prostate cancer that I prefer to show you is the model of CML, and that is that oncogene addiction, which reflects the early aspect of this disease, cannot predict for what you need to do in blast crisis, and vice-versa. So the notion that there are events early that are distinct from events late is already known and experienced in oncology. The difference in prostate cancer is that the model is more difficult. There is a group of patients who need nothing. And then instead of oncogene addiction, there is microenvironment dependence. This microenvironment dependence in this relatively indolent phase -- but clearly progressing down the lethal path -- is a network of signaling pathways that's going to require rational combinations of molecularly targeted agents, and maybe radiopharmaceuticals. And then we have androgen-independent progression as a very late and aggressive event, and I would argue there's increasing evidence to show that it's attributed to alterations in cell cycle.

So this is the model I'd prefer that you think about. There's DHT dependence, and the tissue relation is AR in its normal amount that's present. There's the evidence of transition from endocrine to paracrine, which is the first milestone to the potential lethal progression of this disease, that's manifested by an increase in AR expression or alterations in AR associated with CYP17. There's a paracrine-regulated disease that is more complex, that is going to require combination therapy, and this is characterized by Src overexpression, Rb, amplified and/or altered AR, and increased expression of CYP17. And of particular relevance in this subset is that this is the one that is going to require markers to treat this disease. And finally there are the losses of Rb in the final group, increase in proliferation, alterations in cyclin-dependent kinases, and cell-cycle genes that result in high-proliferation and chemotherapy-sensitive disease. And this model is emerging quickly.

I'm of the view, but accused of being delusional, that we're likely to have models and signatures that can define DHT-dependence, endocrine dependence or AR dependence, and the paracrine/autocrine-late chemotherapy sensitive group, within a year, at the most year and a half. That will leave us with this complicated group in between that is going to need to be dissected much more, that reflects an evolution between the endocrine/paracrine forms of the disease.

I believe that these kinds of approaches that many of us are adopting are going to have a greater impact than getting one more drug, with having a 30% increase, or 3-month increase in survival, without knowing precisely how or how best to combine it with other therapies, to create the magic of combination treatment that has cured our other diseases.

So, the paradigm illustrated at the top of the slide, which is simple and applies to the therapy of other diseases, will be replaced by the paradigm illustrated at the bottom, or one that's similar to this, and will have to be marker-driven. As noted in our discussion earlier, we will have to figure out how to apply immunotherapy in a logical way that complements and enhances the effectiveness of our other therapy, rather than being inhibited by it.

And finally, I'd like to say that this has become a team science. And Tom Beer alluded to this. We're a very unique sort of discipline, and I think we've learned together with sharpening our teeth under the conquest of germ-cell tumors that it requires a real multidisciplinary effort to conquer these diseases. And if you've worked in the most successful environments, you'll know that there is an effective dialogue between disciplines that allows you to marshal all the resources that make this happen. So, thank you for listening to us, and we're going to start to talk about many of the questions that you've asked, so let's talk a little bit to the panel.

Dr De Bono: I guess the one thing that I would say in riposte to one of your statements is that we have looked very hard for patients progressing without a rising PSA, and I would argue that that's actually still very uncommon. So, at present, the jury is out as to whether we genuinely, commonly have disease that truly becomes AR-independent. I would argue that the jury is still out, and watch that space.

Dr Logothetis: Johann is making the argument, which I agree with, that because PSA is not up, there's no androgen independence. And I would argue that there isn't androgen independence in isolation, but androgen independence is emerging within those tumors, and that may be too high of a standard to say that they're androgen-independent, because they don't also have PSA at all.

Dr Beer: I think we're going to soon have the tools to push the envelope on that, not only the combination studies that you suggested and that are getting started, but other agents we heard about earlier and in the translational science session that are active against alternative splice variants; agents that can degrade the AR, which at the end of the day is the final common pathway for AR-mediated signaling. So, you know, there's a bit of a Holy Grail of shutting down AR signaling completely, and I think we're going to either get there or get very close to it, and so we may see a more clear clinical emergence of what you are anticipating.

Dr Logothetis: So the religion that has taken over urologic oncology is: if only we had no AR, would we be great. Is that fair to say?

Dr Beer: I think so.

Dr Logothetis: We're going to be testing it soon. I think the fact that we can grow these tumors so readily now, and they're enriched for AR-negative in xenografts, tells us that they can adapt to non-AR, at least in the late stage when we can grow them; maybe earlier that won't be the case.

We have our first question from the audience: Where does polymorphism screening fit in to determine therapeutic approach? Would anybody like to take that?

Dr Beer: I think I could take a stab at it. Genetic polymorphisms are alterations in our genes that are not thought to be fatal mutations, but are essentially natural variations in our genetic material, and have gotten a lot of attention as potential carriers of risk factors for disease. It's been a bit of a confusing field where a polymorphism is often identified as a potential risk factor, and then a body of literature emerges with half the papers suggesting that that was a true finding, another half not confirming that. So it's an area that we still need to get our hands on. But I think in cancer development, carcinogenesis, and then cancer progression, we're probably thinking more about true mutations, loss of function and gain of function mutations, as well as epigenetic changes that have a more profound effect on gene function than most polymorphisms, which typically are defined as changes that alter the activity or expression of a gene by a relatively modest amount.

Dr Logothetis: So you've got skepticism that the best way to try to find a genetic link to responsiveness or resistance is searching for polymorphisms.

Dr Beer: It's hard to predict what will be the best way for all prostate cancer patients.

Dr Logothetis: So, Tom, if you had to pick between searching polymorphisms, versus some other method, which one would be your preference?

Dr Beer: I would look for translocations and genetic mutations. And I would look for those because those are typically selected for, and therefore perhaps more likely to be important; they are stable and findable. There are many things that change in cancer that are very difficult to measure ex vivo, or in models of human cancer. Gene expression can change in response to stress during a biopsy or sample acquisition procedure, whereas mutations are stably assessable and are proven paradigms.

Dr De Bono: Can I comment? Single nucleotide polymorphisms promised much, but delivered little. And in cancer medicine to date I would say arguably we have very little evidence of their utility. However, mutations in germ line of genes like BRCA, BRCA2 that are frequently missed due perhaps to our inability to gather sufficient family history data should be more closely pursued. Because we now have therapeutics, including PARP inhibitors as well as platinums, that are very effective in patients with prostate cancer who have a BRCA-carrier status in their germ line. So what I would encourage you is if there's a family history in a patient with prostate cancer, particularly in a young patient, do pursue this BRCA2 testing, because these cancers are much more sensitive to the platinums and PARP inhibitors.

Dr Logothetis: I think it's fair to say that these kinds of agnostic searches including polymorphisms still have some challenges before them, and I think they're causing dilemmas right now, but the promise is there.

Dr Beer: I think we're going to get through that pretty quickly though, because even with the grants we're writing today, I'm learning from my laboratory colleagues that it's almost simpler to sequence the whole thing than look for something specific. So I think in the very near future, the concept of a whole-genome cancer sequencing strategy is going to come into the mainstream.

Dr Logothetis: So Tom Beer comes from the school of: if gigabytes of data don't inform, get terabytes?

Dr Beer: Yes. And then know what to do with them.

Dr Logothetis: All right, so the problem with gigabytes is terabytes.

The next question: Empirically, why is bone milieu such an important site for metastases in many hematologic and solid tumor malignancies? Is it an autocrine-like phenomenon in PCa? Should more aggressive use of bone modulators, Src, be used earlier before clinical obvious manifestation of disease?

Let me actually take a stab at that. Many of us believe that prostate cancer is a disease that at least through a significant portion of its natural history adopts bone development pathways in its progression. It actually mimics bone, it becomes bone. And if that's the case, then it's apparent why it would migrate to bone. It's not metastatic like a very hardy cell that is enabled to do everything. It actually can't distinguish the bone from its primary and it goes there. And then you have the limited sort of universe of targets to search for because you look at those critical nodes that are important for bone development to inhibit them, and Src would be present.

You could make an argument, and there are data to suggest it, that, unlike chemotherapy, which targets this autocrine rapidly-proliferating disease at the end, the exact opposite relationship is what we would get with this -- that earlier use of microenvironment-targeting therapies would be more effective before those pathways become more redundant and more complex. If you are prepared to consider paracrine AR-signaling as the primary stromal-epithelial interacting pathway in bone, then evidence that patients with early use of abiraterone appear to be responding better would be nice if some of the neo-adjuvant data that is coming out soon would support that.

Furthermore, while we use sunitinib in the advanced metastatic setting, I can tell you in the preoperative setting there are pathological complete remissions with sunitinib, an agent that has literally no effect in the metastatic setting.

I think ultimately what we're going to have to do, both for Src inhibition and these other pathways, is develop a marker that reflects this biology that will allow the physician the efficient examination of this.

The next question is: can you comment on the concept to modulate the bone microenvironment to wake dormancy of prostate cancer cells? Is anybody willing to talk about issues regarding modulating the microenvironment to sort of create a more hostile environment, if you will, to drive sensitivity to cancer cells?

Dr de Bono: We have no data.

Dr Saad: I don't know this concept of "waking up" the cancer. I would promote more the concept of making the bone a more hostile environment for the tumor to stop it from being able to proliferate and activate the osteoclast rather than waking it up.

Dr Logothetis: I think the notion of shifting the microenvironment of the bone in a way so that it is less hospitable is one that many of us are exploring. Many of us have tried the endothelin receptor antagonists and they have failed, perhaps the drugs are inactive. The denosumab data suggest because of the delay in time to detectable metastases that that's feasible, although the degree of benefit is modest. But it suggests that that's not an unreasonable way of thinking.

So, which of the following are the most important barriers to optimum management of CRPC before and post-test?

So it's different, but not necessarily better, right? Because none of us know what the right answer is. There's an increasing appreciation of the biologic heterogeneity.

Dr Beer: So we've succeeded in presenting how confusing the disease is.

Dr Logothetis: Confusion therapy and speculating is our art form.

So you guys who were previously convinced in hormonal therapy remain that way. Skeptical about chemotherapy, I'm actually surprised. I think chemotherapy is a sleeper and this really works: What do you think about chemotherapy, the sort of skepticism with chemotherapy?

Dr Beer: I've got a bunch of comments on those bars here. I'm shocked to see that hormonal therapy is rated so low. I would argue that AR-targeted therapies are the targeted therapies of prostate cancer of today. Targeted agents are clearly the future, but we've got quite a bit to learn to begin coming close to the clinical benefits we're delivering right now with AR signaling pathway inhibitors. Chemo, I think, just reflects our lack of sophistication in using it. It's clearly a very useful agent, we all know that there is a subset of patients who benefit a great deal, but we still use these agents in everybody, without understanding the subtypes of patients and tumors, and I think once we get smarter about chemotherapy that we'll be able to deliver a lot more benefit with it and hopefully convince our audience of that.

Dr Logothetis: Yes, and the bone-directed therapies and the radiopharmaceuticals?

Dr Beer: The chemo was my part of the talk, so clearly I did a bad job!

Dr Saad: The question here is asking about which of the emerging treatment strategies hold the most promise for the future. I think hormonal therapy is already accepted as being the base of the future, and I think we're going to go to more molecularly targeted agents in the future.

Dr Logothetis: I can see how they could interpret it that way.

Fred, you did a good job!

Dr Beer: It does it all!

Dr Logothetis: Src-family kinases. I think the biology is really at the center of this event. We'll know very soon, you know, I wish the trial well on whether it's positive or not. But the biology is clearly exciting and can be modulated.

Dr Saad: It's always nicer to have a rationale before seeing the results of a trial, than seeing the results and trying to figure out why it worked.

Dr Logothetis: All right, this one's interesting: We'll start from the other end. Fred, do you want to decide on the circulating tumor cells?

We succeeded in creating pessimism.

Dr Saad: I don't know if circulating tumor cells are going to be a true biomarker until we start actually analyzing the individual cells and getting a better idea of what we're dealing with. So, I guess the future isn't circulating tumor cells but in the genotyping and the phenotyping of those cells.

Dr Logothetis: Johann, you're a passionate supporter of the concept. What happened? What went wrong?

Dr de Bono: Well, I think we need more data. I think it's difficult to hypothesize what's going to happen. But I am absolutely convinced that molecular characterization of the whole exome, and RNA expression are going to substantially impact how we manage this disease. I have absolutely no doubt about that. Whether that would be from CTCs, or from biopsies, I'm agnostic.

Dr Logothetis: So you believe marrying those 2 concepts will be critical?

Dr de Bono: I have absolutely no doubt that in 10 years' time we'll be deep-sequencing all our patients for whole exome or maybe 300 to 400 selected genes and that that will drive treatment. And whether that tumor DNA will come from CTCs or tissue, I'm agnostic. It will probably be CTCs for prostate. We'll see.

Dr Beer: I think one of the things about CTCs that's very interesting to me -- and this is speculation, there are no data that I can cite -- is how representative are CTCs of the tumor we're trying to treat. Another question is how representative is a single metastatic site of the global heterogeneity of a particular patient's burden of disease? And I would be curious to see, as CTC technologies improve -- and there are technologies that detect 10 or 20 times more CTCs than the CellSearch platform -- if we find that the CTC sample is actually more of a collective reflection of the extended disease than a single site of biopsy, which may or may not reflect all of the rest of the sites.

Dr Logothetis: We've become enamored with these pictures when we get a biopsy specimen that that biopsy reflects biology better than some test of circulating blood. But what Tom is pointing out, it's only one piece at the other end of that needle. There may be more heterogeneity based on the acquisition issues than is reflected in the cell circulating volume. We share that view and we've got this big effort to understand the circulating microenvironment as we call it, to try to get a handle on that.

So abiraterone acetate inhibits which of the following enzymes?

They did a good job.

So the question is if you had to guess if it was Met, VEGFR2 or both that drive the disease, Johann which one would you pick?

Dr de Bono: I couldn't possibly say, to be honest, I don't think it's just Met. And I don't think it's just VEGFR. But I can say that with potent VEGFR 1, 2, and 3 inhibitors we have seen bona fide sustained, durable responses. And these are drugs that do not impact Met. But I don't know, XL184 doesn't only hit Met and VEGFR2, it also hits multiple other targets -- RET, PFGFR -- so you know I really couldn't possibly speculate what this drug is targeting.

Dr Beer: I'm sort of struck by the fact that everybody knew this before we started talking.

We should have been a little more "higher-end" for this audience.

Dr Logothetis: So the issue is that what may be good for biologists -- to isolate these very specific targets -- may be bad for therapeutic-minded people who want a more broadly inhibited number of targets. The constant dilemma -- we like to know what we're doing, but what we like even better is antitumor activity, so sometime we'll take dirty drugs that are efficacious than clean drugs that are cute.

I'd like to thank you all for being here. I think you've got a flavor of the excitement, and what the future holds for us. The world is going to change and I think, as group, we are convinced that markers are going to drive this and a future is going to be different in 3 to 5 years from now.

So, again, I'd like to thank the audience for taking the time to join us, and a wonderful panel here that I had a privilege to be a part of. Thank you.