Dan George, MD: I want to talk to Chris because he’s got a little skin in this game and has been thinking a lot about this subject. Chris, give us a little background from the clinical trial perspective. Bertrand told us a little about the natural history, biology, and prognosis. What were the data presented at ESMO [European Society for Medical Oncology 2019 Congress], and what else should we be thinking about in terms of this debate, if you will?

Christopher Sweeney, MBBS: Thanks, Dan, for the opportunity to actually go into this. I’m going to take a little time. But actually, what I’ll do is try and address a few issues that have caused some confusion and try to put it all in context. I’m not going to answer any 1 specific question in isolation. I’m going to try to put it all into context. Let’s be very clear: there has been a lot of confusion out there. People have come back to me saying, “I’m so confused about this.” Let’s explain where the confusion comes from.

Two groups of people put together 2 trials, which were similar with the same drugs. One had a very inclusive approach, the STAMPEDE trial team, and they interpreted their data with a certain analysis plan. We on the CHAARTED trial side did that analysis plan with very predefined subgroups. The STAMPEDE team said, “We don’t see a difference between M0 and M1 in the end points in the early analysis. Therefore, our conclusion is the drug works across the board.” On the other hand, we on the US side said, “We don’t see a difference between the M1 specific group with a low burden and a high burden that was prospectively designed.” That’s where the confusion comes from.

Right now, where we are in 2019 in the data that got presented today at this meeting at ESMO 2019, is 2-fold. One is the updates on the N0 state, which is both a rise in PSA [prostate-specific antigen] postprostatectomy, like the EMBARK trial. Then for the adjuvant state, like the bolus study, 82 radiation plus or minus docetaxel. Our colleague, Joe O’Sullivan, has been an active member in that STAMPEDE team, so we’ve got the whole perspective here represented.

We also have a French study, which is a third analysis. That’s why I also say, “Let’s look at all the data sets, see what the patterns are, pick up, and look for reproducibility and consistency.” That’s my first message.

Dan George, MD: Across studies, yes.

Christopher Sweeney, MBBS: Across studies, yes. I think it’s like listening to the radio station, and we’re finally tuning in to the band we want to hear. I just want to go through to Bertrand. There are 4 disease states with different prognoses in the hormone-sensitive setting. State 1 is de novo high volume with a median survival of 3 years in all the studies with testosterone suppression alone.

Dan George, MD: Defined as visceral or multiple bone METS [metastases] greater than 4, or 4 or more.

Christopher Sweeney, MBBS: Correct. Most patients who present with de novo disease go to their physician, their primary care, with back pain and a PSA that’s 300. They have 4-course imaging, and it’s problematic. Then you’ve got the other extreme, which is the patient has a rise in PSA after their prostatectomy or radiation, and the decision has been made, “Let’s image you and stop therapy at the time of all imaging seen on the CT [computed tomography] scan.” That median survival from the time of ADT [androgen deprivation therapy] is 8 years. We’ve got 2 bookends, and we’re putting them into the same study.

I believe they are very prognostic, and the question is, are they predictive for therapy? We’ve got this middle disease state, which is the de novo low volume, and when we look at all the patients who come through Dana-Farber Cancer Institute, and we have about 500, this accounts for 20% of the patients with de novo low volume because it’s unusual to present with a little PSA and get a scan done. About 50% of those patients are actually probably chemotherapy fit. It’s 10% of the pie that was presented and clarified at this meeting. So just put that into perspective.

The other rare entity is the patient who has rapidly progressive disease with high-volume disease after prostatectomy or radiation. That’s probably these patients who had surgery just 6 months down the road from this group.

So they’re the patients we’re analyzing. That’s the first message when you sit down with a patient: let’s actually work at who you are based on your chemotherapy level of fitness, your volume of disease, and your metachronous or synchronous de novo, your prognostic factors, and your comorbidities. That’s the first thing we should do whenever a patient sits with us.

Then we can have an individualized precision medicine approach based on the clinical factors about what’s next. Now let’s look at our tools and the data sets we have for our tools. Our tools are the new potent hormonal therapies—abiraterone, ...., enzalutamide, and apalutamide we have data with, and docetaxel. Let’s not forget about our friend radiation to the prostate in the de novo and metastatic patient.

Then we have a nuanced conversation. I first of all ask the questions, “Does the patient have de novo metastatic disease? What are my options? Are they chemo fit?” Now let’s look at all the data sets. All data sets, GETUG-AFU 15, the STAMPEDE updated by high volume, and CHAARTED all show a treatment effect in the high-volume de novo metastatic disease. How great that is varies between the studies, but it’s all in the same direction—no question.

Where there is some variation is the de novo low volume, and the STAMPEDE team presented that today at this meeting, #ESMO2019. So we see that the treatment effect in their group was about 0.76, which is about 300 patients in that group. When we look at the treatment effect in the de novo low volume with about 150 patients in the CHAARTED study, it was about 0.86, going in the same direction. I always said there’s a subgroup in this, but I just don’t know who they are. We don’t have enough. When we look at the GETUG-AFU 15, they’re about 0.9 or 1.0. So we don’t see anything in that group.

But it’s 2 out of 3. As Meat Loaf said, that ain’t bad. There’s some consistency, but it’s not as robust as the high volume. Now let’s go to the “not chemotherapy fit.” Chemotherapy is clearly off the option. But you’ve got radiation for prostate, apalutamide, enzalutamide, and abiraterone all as options. You choose. We have to have the patient involved in that conversation.

The next question is, what about the patient who has relapsed post local therapy? The only data sets we had in that group is the —amides: apalutamide and enzalutamide and their forest plots in the 2 New England Journal of Medicine papers showing a treatment benefit with a hazard ratio of about 0.5 in those groups. So we’re dialing in on that.

I do want to come to 1 other very important message where people would come up to me with confusion saying, “Well, what about the N0 state?” I’m going to push and say, “I still think docetaxel is predictive for benefit for the patients with a higher burden, and it’s de novo.” This is because we don’t see any benefit in the relapsed setting. But the other low-volume relapsed group where we don’t see a benefit in 2 studies is with patients with the N0 docetaxel, rising PSA patients. So the patient has rising PSA, post prostatectomy. There is a study completely negative for MFS [metastasis-free survival] and OS [overall survival]. The STAMPEDE group presented, and they didn’t see a benefit in that group. That low-volume relapse seen with 2 bone metastases on a CT scan and rising PSA is the same group, and it’s predictive for not benefiting in that.

Then there’s another conversation about the adjuvant study N0, which is another group altogether. The RTOG 0521, GETUG-AFU 12, and the STAMPEDE studies just presented on this. There’s a slight benefit in some patients. We need to get that data set into 1 individual past-patient data meta-analysis and work at who that subgroup is.

Transcript Edited for Clarity