Are There Predictors for Successful Treatment of Oligometastatic Prostate Cancer? APCCC 2022 Presentation - Piet Ost

September 22, 2022

At the 2022 Advanced Prostate Cancer Consensus Conference (APCCC), Piet Ost presents predictors for the successful treatment of patients with oligometastatic and oligoprogressive prostate cancer,

Biographies:

Piet Ost, MD, PhD, associate professor at Ghent University, Ghent, Belgium, and radiation oncologist at the Iridium Network, Antwerp, Belgium


Read the Full Video Transcript

Mary Ellen Taplin: Next we have Dr. Piet Ost, who will talk to us about, are there predictors for successful treatment of oligometastatic prostate cancer?

Piet Ost: Thank you. So a couple of weeks ago, I asked if this truly was my title, because we all know that there's not a lot of evidence, let alone we can predict who benefits. So that was one, a big issue for me. But nevertheless, I'll give it a shot. But first, I have to give you some history and I think that's extremely important. When we started more than a decade ago, doing metastasis directed therapy, this was the situation. So me, I'm sitting on the ground trying to pull the train out of the station, using my teeth, using all my strength. However, there were several urologists, some you might know, that said, hold on, you are going crazy. Next step was you're going Pokemon hunting. This is not to a way forward. So I was really trying to get that train moving from the station.

Now we are a decade later. Now, urologists are still on the train, but they're going full throttle. Now in between 10 years ago and now not that much evidence actually was generated. But nevertheless, now I am trying to catch up to get to that train. But maybe we're going too fast. And the big question is which patients do actually benefit and that's my topic. So let's see if we can find some.

So, first of all, I think it's very important to differentiate between prognostic and predictive. This is an artificial construction, but just see that the black lines are biomarker positive, the gray lines are biomarker negative experimental treatments are solid lines. So here you can see that the biomarker is prognostic, meaning that if the biomarker is positive, patients do better irrespective of the treatment they get. So that's prognostic. Now this biomarker is predictive. Here you can clearly see that for the black lines, if patients get the experimental treatment, being the solid line, they do better. For patients in gray, biomarker negative, either they get standard or experimental treatment, there's no difference. So this here is a patient is biomarker positive, they should get a new treatment. It's clearly predictive.

Now, often we are in a different situation. This is when the biomarker is predictive of a differential benefit, and that's pretty important. You can see that a patient in black who is biomarker positive, benefits even more from the experimental treatment as compared to the biomarker negative. But even the biomarker negative still benefit, but to a lesser extent. And that is important to set the scene a bit.

Let's start with synchronous. Of course, there is no evidence, but I do want to show you one very important predictive biomarker. And that is imaging. And of course we only have standard imaging in the big trials, but here, which patient needs treatment to the primary? It's pretty easy. Low volume has a benefit on overall survival. High volume, no. So there's a clear benefit here for those patients and the biomarker is counting metastasis. So counting metastasis is predictive of the outcome of local radiotherapy. So that's an easy biomarker, and that's why we need to consider imaging as well as a potential biomarker. And probably we will do this in the future as we've seen with PSMA PET measuring volume. No longer counting, which might be even easier. Are there molecular data as potential predictors? Well we looked in the literature to see if there are molecular differences between high volume versus low volume. And indeed there are. There are important reported differences in some molecular aberrations. For example, patients that have TP53 mutations, BRCA, they have a higher likelihood of having high volume disease when you diagnose them. Of course, it doesn't mean this is predictive. No, this is prognostic information, but maybe we can use this in the future.

And the way we use this will probably be biological guided treatment. If you have the whole pool of patients, you look at the primary, you look at CTDNA, whatever marker you think, you put that in whatever machine you want and the output is a randomized trial, which randomizes the standard of care versus a treatment depending on the biomarker you see. And there's several efforts now doing that. Just pulling a plug for ProBIO, because I'm running this one in Belgium together with the colleagues in Sweden. And just open in Switzerland I heard. So that's very good news.

What about metachronous? So those are the patients that are recurrent. This is where most evidence is. Imaging as a predictor. Do we have one? Actually, yes. And it was already in 1995 that [inaudible] said the limited effectiveness of these treatments... And for example SBRT, it's because we don't see them. It's the inability to recognize all metastasis. We see the tip of the iceberg. And a beautiful example of that was done in the Oriole trial. So here you see ORIOLE randomized SBRT versus observation. Everybody got a PET, but everybody was actually blinded. So as you can see the blue lines, those were patients that had a PETs and conventional imaging showing the same extent of disease, the same lesions. So they got SBRT to all lesions visible on both modalities. The line in red is where PSMA saw more lesions, but of course the physician was blinded. So they could only treat the conventionally imaged lesions. As you can see, time to distant metastasis is worse. So if you want to do SBRT, metastasis directed therapy, use the most sensitive imaging you have available. And probably PSMA is an interesting predictor of outcome.

So what about molecular data as a predictor in this specific setting? There are probably molecular differences here. De novo versus Recurrent. Yes, there are as well differences in molecular aberrations. The question is, do they help us? But even within a disease state, there are differences. This was a study we just did where we looked at the primary. All patients got radical prostatectomy, but they recurred in the lung, which is not that often that it happens, less than 5%. But nevertheless, we tried to get as many lesions out of there as we could, because they were diagnosed on coal lines. So we were unsure what they were. Then we went for sequencing. And as you can see, it is clear that there are differences between the reported literature. Patients that develop lung mass do better. They don't have TP53 aberration. They have P10, but all the other bad ones they don't have.

And all these patients in this cohort actually are still alive at 10 years follow up. So here clearly, molecular information can help us and lung patients with lung recurrences should not be excluded from trials because they do very, very well. So not all visceral METs are the same, so that's of interest. So the big question is, is there a subset of patients that does not benefit from SBRT or observation? So what we did is we pulled ORIOLE and STOMP together. Both trials did the same as SBRT versus observation, targeted sequencing of the primary or the blood. We looked at high risk mutations, and this was an arbitrary definition. So we looked at pathogenic or somatic or germline mutations in ATM BRCA, RbA and TP53. But the hypothesis was, patients with a high risk mutation do not benefit from our intervention. That was a hypothesis, prime endpoint PFS.

So for patients with no high risk mutation, blue line is observation, yellow one is MDT. You can clearly see that SBRT is better than the observation, but patients do quite well. So if you have no high risk mutation, outcomes are okay. But do mind that five year almost everyone is relapsing. What about patients with a high risk mutation? Well, still there is an enormous benefit of SBRT, but look at the time to recurrence. Patients who are observed with a high risk mutation, bad idea, they recur within six months. Okay, with SBRT, you can increase it maybe to 12 months, but still all these patients are recurring very fast. So that means that MDT is always better than observation, but patients with a high risk mutation apparently benefit even more.

And there was a differential benefit, the largest benefit of SBRT was actually in patients with high risk mutation. But nevertheless, they failed very fast. So that was for us then the next step is, do these patients with a high risk mutation, maybe they need systemic therapy. And that when we compare all patients getting SBRT, no high risk mutation versus high risk mutation, it is clear that a proportion of patients relapse within two years and actually almost all of them. There is a small tale that do better for a long time, but this is why I think that the next step is adding systemic therapy. Probably temporary, but the time of giving SBRT only probably that train has left the station, we should probably add something systemically.

So for me to conclude, I think, what are the predictors of metastasis directed therapy? Well, in De novo, it's pretty simple, use the definition of low burden counter METs and then you know that you need to treat the primary. On metastasis directed therapy, there are no data, so of course there are no predictors. For recurrent mHSPC, use the most sensitive imaging you have. Today that is PSMA. The genomic profile is of interest. Look at the primary because it's predictive of a differential benefit. But if you find the high risk mutation, don't just give SBRT add something, and that will probably benefit the patient. Thank you.