Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: We’ve had, Johann, some very nice discussion on DNA damage in metastatic prostate cancer. I think you covered very nicely the role of PARP inhibitors in that space. If you were looking at the next generation of AR-targeted therapies, what do you think are the most promising of those agents?

Johann de Bono, PhD, MB, ChB: I think what we need now are drugs that are targeting the androgen receptor splice variants. There have been several drugs that are still targeting the ligand binding domain, but I think these really are, for me, not the urgent priority. What we now urgently need are drugs either inhibiting the spliceosome—blocking the generation of AR splice variants—or targeting the AR amino-terminus. And there are trials running in the clinic now with these kinds of drugs. For me, these are the priorities, so let’s wait and see.

Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: What about the use of checkpoint blockade in prostate cancer?

Johann de Bono, PhD, MB, ChB: In the United States, we already have pembrolizumab approved for mismatch repair—defective cancers of any type. Therefore, in the United States, based on nonrandomized trial data, pembrolizumab is being used for mismatch repair–defective prostate cancer. We obviously don’t have data showing survival benefit, but I can tell you from my experience, I know that patients with mismatch repair prostate cancer can have spectacular responses in the end-stage setting with immunotherapy. We have a lot of genomic and immune data on this population of patients. The mismatch repair prostate cancers have much higher PD-L1 expression. They have more immune cell infiltration. It is likely that at least half of these MR prostate cancers will respond to immunotherapy. It does, however, concern me that a lot of these trials for these immunotherapies are not pursuing patient selection, and therefore I would postulate that many of these randomized trials without patient selection will fail.

Chris Parker, MD, FRCR, MRCP: What data do we have on the activity of checkpoint inhibitors in patients who we know don’t have mismatch repair defects?

Johann de Bono, PhD, MB, ChB: Not a lot of data, Chris. One of the challenges for us is that the current assays we have in the lab for mismatch repair come to us from colorectal cancer. They’re not perfect assays. We have a lot of data on this. We have a paper being reviewed saying that, actually, we need to really optimize the assays to detect these genomic aberrations in prostate cancer and indeed in other diseases. To be honest, at present, all I would say is that outside mismatch repair, there are a huge number of questions being asked, and we are running trials to ask those questions. We’ll have those data hopefully in the next 6 months to a year in the public domain.

Joe O’Sullivan, MD, FRCPI, FFRRCSI, FRCR: Bertrand, you mentioned earlier PSMA (prostate-specific membrane antigen) PET imaging as a potential diagnostic tool in prostate cancer. There has been a lot of interest recently in potentially having therapeutic radioisotopes attached to PSMA-targeted therapy. What’s your view on where we are with that? I think Lutetium-177 is probably the most widely used, but there are also some alpha-targeted therapies.

Bertrand Tombal, MD, PhD: There is opportunity and there is threat. The opportunity is that, clearly, there is a rationale. PSMA is good to target PSMA-expressing cells. Lutetium nuclides, as you know, have high local activity, but the threat is that it’s easy to make. My impression, at least in Europe, is that we have confronted a large group of physicians who believe we don’t need a randomized controlled trial or the comparative data to use it. And then, when you look at the evidence, what you see is a PSA decrease. You see a change on PET PSMA. You use the target, and you show you hit the target. I think it was nicely pointed out, by Karim Fizazi, that our impression is the response rate is very long. There is some toxicity, so I clearly see that we need randomized data.

I was telling Johann, I remember the early stage of the history of cabozantinib in prostate cancer. Everybody was amazed by these bone scans turning to 0, but it didn’t translate into significant benefit. It’s select. The nice thing is that you have a nice precision medicine process because you select the patient based on imaging. So, that’s a good thing. You take patients enriched in PSMA. We have almost finished a trial where every patient had whole-body MRI and PET PSMA scans. We can show you that almost every patient who is PET-PSMA positive will have deposit on the whole-body MRI, which doesn’t express PET PSMA.

Now, we don’t know all of the responses to the therapy. That’s the first observation. Then, the drug is over selective, and to me, there’s still a bias in the measurement of the response. So, the effort should be to run the proper trial because, otherwise, we’re never going to get the answer.

Chris Parker, MD, FRCR, MRCP: Of course, I agree with you, Bertrand, that we need to run the proper trials. There are all sorts of unanswered questions. It’s not just which isotope to use. It’s also which targeting vehicle we use, what dose, what interval, and so on. So, there are loads of unanswered questions. But I am much more optimistic than you sound with regards to the end game, because I’m quite sure it’s not another cabozantinib, because the responses that we’re seeing are not just imaging responses. We’re seeing patients who are sick becoming well, really dramatically. I’m sure it’s an active population.

Johann de Bono, PhD, MB, ChB: I would agree with you, Chris. I think particularly for the alpha particle methods, more than the beta particle methods, we already know that radium works, and we already know that we have antibody-targeting methods that will deliver this DNA double-strand break-generating compound in the tumor cells. I, like you, am very confident these drugs will make a survival difference. But on the other hand, I do agree with Bertrand that there are many concerns, not least of which is heterogeneous PSMA expression, downregulation of PSMA expression, if we’re targeting that target after drug exposure. So, like you said, we just have to get the trials done. The concern I think we all have is that the hype around this area needs to be consolidated by robust data with imaging that is not PSMA PET. I think that is a major weakness of what we’ve seen to date.

Chris Parker, MD, FRCR, MRCP: It was interesting, because we had data presented from Australia yesterday, a series of 30 patients treated with PSMA lutetium. They presented the PSMA responses, and they presented the PSA responses. In the small print, there were some RECIST responses. Now, the RECIST responses looked really impressive—10 out of 17—but that was almost glossed over; whereas, to me, that’s the most important efficacy data of all.

Johann de Bono, PhD, MB, ChB: I guess that’s nuclear medicine physicians and oncologists presenting the data. That’s really essential.

Bertrand Tombal, MD, PhD: But once again, it’s very important to induce sustained response.

Johann de Bono, PhD, MB, ChB: Indeed.

Bertrand Tombal, MD, PhD: And that’s one of the problems. Are we sure? Because if the response is not sustained, we’re going to have to repeat and repeat and repeat and see toxicity. So, once again, we fully agree there are individuals who obtain great benefit from that treatment. We’ve all seen patients like this. But the question is, if you apply it to a large number of patients, do you get enough benefit to justify the single benefit in the few individuals?

Johann de Bono, PhD, MB, ChB: What I would make very clear is that a lot of the work that’s been done and presented has actually involved—particularly from the Heidelberg group—the small molecule enzyme inhibitor bound to the alpha particle emitter, actinium. I think the antibodies are going to be quite different. I am concerned that a lot of the work so far has been done with antibodies that have not been manufactured with the GLP (Good Laboratory Practice) regulatory requirements that we’re all used to. I think, while these data are very interesting, as a drug developer and as you’ve all echoed, we need a lot more carefully designed trials to ask some really critical questions about these drugs.

Bertrand Tombal, MD, PhD: But that’s why the success of a drug is not only the activity of the drug but also the development plan.

Johann de Bono, PhD, MB, ChB: Correct.

Bertrand Tombal, MD, PhD: We have seen beautiful drugs destroyed by the development plan. And that’s why it’s so important that we need to stick to the fact that if you don’t comply to a strict development plan, the first thing you hurt is the drug itself. We know drugs like satraplatin, where if you had to repeat it, we probably could repeat the trial and obtain a benefit from the drug.

Johann de Bono, PhD, MB, ChB: And there’s no doubt that if you rush drug development, you’ll spoil the cake, you’ll overcook it.

Transcript Edited for Clarity