Read the Full Video Transcript

Howard Scher: It's a pleasure to be here, and thank you for the invitation. I'm going to try and approach this topic both from a point of view of a clinician as well as a researcher. So here are my disclosures.

The first thing, that first question I asked is, "What is this entity?" And I think as we heard from the last few days, it's changing quite dramatically in a relatively short period of time. If we think about clinical states, you can follow the patients from localized disease to the biochemical recurrence. Essentially we are still using the standard FDA definition of castration, but essentially it's a patient who has no detectable disease by a radionuclide bone scan. Different groups will also include MRI but also include CT. But you have to keep in mind that this is a population that's generally asymptomatic, with the exception of some of the side effects they may have had from therapies.

Here is one of a recently published, really outstanding review on the topic, from Dr. Mateo, published very recently. And is starting to get a sense of how heterogeneous this population can be. And as we're using the drugs in advanced disease earlier, it's going to get even more heterogeneous, which is a factor that we have to think about going forward.

So if you trace the patient from their first diagnosis, those who were treated with curative intent to ultimately relapse with a biochemical recurrence, some are observed for a period of time. Many are ultimately treated based on risk. And only those patients who have never had any detectable disease on metastatic, on imaging, are the ones who would fit the definition, if you will, of non-metastatic castration-resistant disease.

So the next question is, who needs treatment? And what you're doing here is trying to balance prognosis versus competing risks of death. Because in many cases, these are older patients, some of who can be quite frail.

A number of studies have addressed this question. Really the first to do it systematically was Dr. Smith from Boston where he looked at cohorts of patients and was able to identify probabilities of patients developing metastatic disease based on PSA doubling times. In his Phase III trial of denosumab versus placebo in which many members of the audience who are also investigators on, they defined high risk as patients who have a PSA of eight or greater and a doubling time of less than 12 months. This was a drug that only targeted bones so the primary endpoint was bone metastasis-free survival and they were looking for seeing a 20% reduction or difference between the two cohorts in terms of that clinical event.

And are looking in and working in several series, how often is this high risk occur? It's about a third of patients and this is based on a work by Dr. Antonarakis at Johns Hopkins. So it gives you a sense of the frequency of the patients who are actually going into studies. This trial ultimately showed an improvement in the bone metastasis-free survival, but in an ODAC, which you'll hear more about from Dr. Hussain, the 4.2 months difference was really only about one imaging interval and was felt by the committee, just not enough to justify an approval, particularly considering some of the adverse events that were encountered.

And is the case with most of the studies, followup is too short in order to see a survival difference. So again, OS was not prolonged, but if you count the number of events, it's relatively few and I would argue that looking for overall survival is going to become meaningless because of post-treatment and therapies that are effective.

Now when you're looking at trials going forward, you're trying to balance the need for treatment in this asymptomatic population versus the potential adverse events that occur from the drug that's administered. And most widely used is PSA doubling time. And now more contemporary trials are looking at metastasis-free survival which can include soft tissue disease, which a bone targeting agent would not be expected to affect.

If you start looking at where this come from, this was again looking at Matthew Smith's study where he looked at the event rate based on doubling times, you can see on the left. But what was particularly interesting was the inflection point and somewhere at a doubling time around eight months, you can see a rapid increase in the risk of developing metastatic disease. And you find most of them were contemporary trials are using this timeframe for enrollment.

But you also have to consider competing risks of death. This was not a particularly large series, but it gives you some of the issues that have to be thought about when you're thinking of your trials and who to treat. If we look at the patients with rapidly doubling time, you could see that the younger patients have a higher risk of a cancer-related death on the left versus those on the right. In contrast, once you start getting into the slower doubling times, the risk of mortality is much higher from other causes.

And again, the factors that were considered here were age, comorbidities and doubling time in looking at these estimates. And this is the type of new information that we have to integrate into our decision making, not only in terms of treatment but also in terms of trial design. And I'll show you a few other examples of nontraditional types of reporting that are beginning to be presented. You heard a very nice talk from dr Sykes looking in three dimensions. This is not something we typically do and again, these are all biomarkers in their own right that are going to have to be validated in terms of their significance going forward.

So think about therapeutic options. The world has really changed. In the past, we were looking at bone targeting agents, endothelin antagonists, agents that inhibited vascularity such as bevacizumab, none of which actually were successful. But as we got into the next generation AR directed therapies, we see essentially three positive trials. Abiraterone, just to mention has actually not been evaluated in a definitive Phase 3 study, but has very, I'll say compelling data in Phase 2.

So if you just look at this at a high level, this will be detailed by the following speakers and when you look at the number of patients in each of these trials, PROSPER looking at enzalutamide, almost 1500 patients, SPARTAN, 1200 patients, ARAMIS, 1500 patients. These are big costly trials. All of them hit their primary endpoint, were successful and again, not surprisingly, no improvement in survival. You would not expect them because these events occur significantly early. And you can see the marked difference in terms of the endpoint, which is bone metastasis but which is metastasis-free survival was quite significantly different between the two treatment arms.

I was very skeptical that these trials could actually be conducted. You're using AR directed therapies which you know lower PSA almost immediately if they're going to work. And if you look at the left, you can see how quickly a number of the patients progressed and developed a disease. But to me, it was very impressive that these patients could be maintained on study knowing that their PSA progression had occurred until the metastasis event. And that easily could have been skewed if those control arm patients took abiraterone for example or another approved drugged. And you look on the right in terms of the primary endpoint, again looking visually, this is a clear separation. You don't need a caliber to assess.

But now we're starting to see different kinds of data. Here's a radar plot trying to factor in individual adverse events that occurred on study. Again, I don't think this is really necessarily a direct comparison, but this is how we're trying to balance the risk-reward ratio, which is key for any indication and ultimate use. This is what we think about. We know we have choices. What are the factors that we consider most when we say this particular patient should get this particular drug?

So again, a new type of presentation of information. While you may see things that look significant, again, it's not clear how to interpret these as yet just by looking visually. We have to validate their significance going forward.

I'd like to move now to what I consider the times they are changing. Now we have nine drugs or more. I probably should stop counting. It's going into double digits very shortly. I started in this field when there was essentially nothing, and what's happening now is all of these agents are being used early. So the biology of the tumor that we're seeing in the non castrate metastatic space is not the same. You saw the genomic profiles that Wassim Abida had developed and you could see the difference in terms of the biologic effects, but we're also seeing the Will Rogers effect. Why am I showing a cowboy? Because the actuals, we're seeing a difference in the stages. Here's a PSMA pet scan. We showed a lot of these yesterday from Michael Morris, and we have a very aggressive interventional radiology group. We want tumor and this was a histologically confirmed, a metastatic lesion. You could see it was quite small.

We also know that you'll see more when the PSA is higher, but how do we present this? And this is an example of a new way, a staging system if you will, the molecular imaging TNM that was reported by Dr. Eiber. And you can see there are different patterns of progression. And in some cases you have inter pelvic, some patients extra pelvic. So if you think about what's going on, there was a series that was a combined of eight thousand plus patients from different centers. And in that cohort happened to be 200. I was asking around to my colleagues who I will thank shortly.

Is that a double or a single? Huh? It said, "Give me real quick." All right. So we see that there's a lot of patients again who have only recurrence in the pelvis. They showed an interobserver agreement. They validated in different methods and clinical management was recorded.

Just to conclude, we're starting to see new measurements. We have to focus on what's the decision that we want to make with the tests. We have to validate the methodology. This includes how the images acquired, interpreted and reported. We have to show that it means something the same way we do with a drug for an indication. And just because you change management is not clinical utility.

The FDA now recognizes this as an endpoint. We have options we did not have. PSMA is changing practice, but we have to use it more intelligently and show that it impacts on patient management if we want to use it. And this just shows the importance of shared decision making. Just to conclude, I want to thank Doctors Eiber, Fendler and Fanti who helped me in terms of getting data to present. In particular, on some, that's just in the press as of Wednesday. Thank you.