Pedro Barata: Thank you so much, Neeraj, for giving me the chance to sit down with you and chat about prostate cancer.
Neeraj Agarwal: I'd like to talk about your studies. You have done a lot of work in testing in the mCRPC setting, especially from the perspective of homologous recombination repair mutation testing, and I think this can be extrapolated to other genomic testing, too.
So, first of all, I'd like to talk about the JAMA Oncology paper you published, where you talked about testing in the real world, based on, I think, Flatiron. So, I would like to ask you about the results of that manuscript you published in JAMA Oncology, and then we can talk about the implications in the real world and on our practice.
Pedro Barata: Sure. So, I think none of what I'm going to say is going to be very surprising to you, because you also look at research questions in different datasets as well. So, not surprisingly, we know that patients in the US are being under-tested, right? So we see that less than half of the patients diagnosed with metastatic prostate cancer, and it was also applicable to mCRPC, are being offered any form of genetic testing.
Not even to get into the details, but whether you include germline or somatic testing when you look specifically at HRR. And why HRR testing? Well, because we actually already know the prognostic value, particularly with BRCA. But also the predictive value with PARP inhibitors. We know doctors change their practice based on treatments available to patients, as has been the case across really all of oncology.
So, we looked in a more systematic manner into the Flatiron database, as you said. We basically were able to see that, number one, less than half of the people are being offered testing. Testing happens at different time points. More often than not, it happens at the time of CRPC, but also after one or two lines of therapy.
You also have a number of patients being tested very late in the game, for which, and that's my prediction, many of them end up not getting any systemic therapy afterwards. So, I call it almost a desperation move. In other words, I've exhausted—
Neeraj Agarwal: So, desperation testing, if you will. So patients are going to hospice, and they are getting tested three or four months before going to hospice.
Pedro Barata: Right.
Neeraj Agarwal: Those test results don't really help those patients.
Pedro Barata: No. Because the turnaround time is not a week, right? It's usually several weeks, and you've exhausted life-prolonging therapies, you don't know what else to do, you're ordering that test. A lot of times patients are declining functionally and end up going to comfort care. You're not able to act on the results of genetic testing.
So, that's in regards to HRR. Now, you can extrapolate and say that's actually true beyond HRR. We have other biomarkers that we could be focusing on. Beyond HRR, you could think about microsatellite instability-high tumors, for example. 2% to 3% of patients with metastatic prostate cancer for which immunotherapy is very appropriate and should be strongly considered.
Neeraj Agarwal: And TMB-high status.
Pedro Barata: And high tumor mutational burden is the other aspect that we should think about. But not just that, right? Even for chemotherapy, we came up with this molecular definition of aggressive variant disease, with TP53, RB1, and PTEN—these are tumor suppressor genes.
When you have the presence of two out of those three, you know those patients are destined not to do well. But also, you know that they tend to benefit from intensified chemotherapy. This is data from the MD Anderson Group, Ana Aparicio, Paul Corn, and others.
Neeraj Agarwal: The combination of carboplatin with taxane.
Pedro Barata: Exactly. And so, I actually do that in my practice. So, that's just a few examples. I could keep going and think about PTEN loss, which we could do by immunohistochemistry. We can also look at the gene expression, or the lack of gene expression of PTEN as a surrogate for poor prognosis, as well as an independent poor prognostic factor.
Neeraj Agarwal: So, let me interrupt you for a second. If we have life-prolonging drugs in these patients, whether it is olaparib or the combination of enzalutamide plus talazoparib, both combinations are associated with increased overall survival.
We have this highly well-tolerated immunotherapy drug, pembrolizumab, approved for these patients. And then other options which can be available, based on if they have AR-indifferent prostate cancer, why are less than half of the patients in your study having NGS testing?
Pedro Barata: Sure. So, we didn't ask the question about why didn't patients get tested. What we know, from looking at different datasets and different studies looking into that, there are a number of barriers to testing. And so, one is obviously availability of tissue. I think liquid biopsies are available now and we can talk about that, and how the concordance rate is pretty high, compared to tissue-based testing.
But the fact is, most of us still rely on archival tissue from the radical prostatectomy, or from a biopsy. That means you have to have tissue available. We usually use a cutoff of about five years to assess the quality of the tissue, because we need to have viable DNA in the archival tissue to be able to sequence that tumor. Right? And that's not always the case.
These patients, at times, recur many years later. You always ask yourself, "Is the disease that I diagnosed initially 7, 8, 9 years ago, the same as the one that I'm facing right now that I have to treat?" So, this is an important barrier, and just one of many.
There are several others, Neeraj. Lack of awareness. If you think about what you learn in medical school about genetic testing, and what is out there, it's not the same as what we know today. Things have evolved significantly. So, awareness and knowledge of the providers.
Neeraj Agarwal: And the burden, extra burden of work on our nursing staff.
Pedro Barata: Logistics, right? We figured that out in Cleveland. I was at Tulane prior to that, and Oliver Sartor had a fantastic model that we tried to implement in Cleveland, by having someone with us, helping us to execute the testing. Because you are right. It's a big burden, because you need to go into the room, sometimes we need the consent.
When you think about liquid biopsy or germline testing, that means you need to send someone to the lab. You need to be comfortable doing the pre-test counseling. Then you want to make sure in the lab they do it with the right kit, you send it out, because for the most part, it is outsourced. Then you need to track the portal to get those results. You need to download the results. You need to explain to the patient what that means.
And you still need to use that, and think, when is the right time to do therapy based on those results? You see how complex that is? So the teams out there, taking care of patients with prostate cancer, need to adapt. It's a big change. So, we're working in Cleveland to try to change that model, into a model that allows us to test more.
We're working—in part, we're really bringing together Case Western Reserve with the Cleveland Clinic for this. We have a project going on which hopefully will impact the community significantly as well. But the point of our conversation is to say, there are absolutely a number of barriers that are not easily addressed. We've been trying hard, but it is a hard thing to address, and it takes time.
Neeraj Agarwal: But one way to address it, at least as far as the quantity or quality of tumor tissue is concerned, is to get ctDNA testing.
Pedro Barata: Right.
Neeraj Agarwal: And you have shown, if I'm not mistaken, in the Journal of Immunotherapy of Cancer, that patients with prostate cancer with MSI-high status had great responses to pembrolizumab.
Pedro Barata: Yes.
Neeraj Agarwal: I think Dr. Oliver Sartor was the senior author. So, we already have data, and then we see the PARP inhibitor studies. CtDNA is as good, if not better, than tissue testing in picking up most of these mutations.
Pedro Barata: Right. So, great point. So, just to summarize, in liquid biopsy, we have the ability to test for MSI-high as well as TMB. We have that reported. I think it's been available for the last five years at least, now, so we can do that. Yeah. So, we put together a nice cohort of patients, all MSI-high, based on liquid biopsy. We also had the tissue for most of them; the concordance is high. And they were basically treated with immunotherapy, the immune checkpoint inhibitor pembrolizumab.
The way I summarize data around immunotherapy for prostate cancer is that about half of the patients are going to have a response, by PSA, but also on scans. So, it works half of the time. However, you need to test 100 patients to find your two or three patients that will benefit.
Neeraj Agarwal: With the PARP inhibitors, the numbers are much higher.
Pedro Barata: The numbers are much higher.
Neeraj Agarwal: I think liquid testing can easily complement tumor tissue testing, especially when tumor tissue is not there.
Pedro Barata: Right. And just to add to that point, two comments on that. So we also looked at TMB in liquid biopsy without MSI. And actually, regarding the cutoff, because doctors use the same cutoff, right? "Oh. Well, let's use 10 mutations per megabase." In fact, we failed to show a benefit of using that TMB in liquid biopsy as a predictor of response to immunotherapy, outside of MSI-high.
So, for TMB-high, the data does not support using liquid biopsy alone without tissue. MSI-high or TMB-high in tissue? That is a predictor for immunotherapy. So, it's informative in a negative way regarding liquid biopsy. When you go to HRR, then you jump from the 2% to 3%, to now up to a quarter of the patients will harbor that.
Now, of course, if you test four patients, the chances are you'll find one. And so, don't get discouraged, because you need to test several patients to find your positive results. If you're only looking for BRCA, that won't be 25%. So, the 25% or so—we're talking mCRPC, germline plus somatic, and we're looking for BRCA, ATM, FANCA, et cetera, right? The entire HRR gene family.
And that has two therapeutic implications, whether you're going to think about a PARP inhibitor monotherapy, or you're going to think about a PARP inhibitor combination with an ARPI. But certainly, that increases the chance of opening the door to a new set of tools available in your toolbox for your patients with metastatic disease.
Neeraj Agarwal: Those are great points. So, just to summarize for our audience today, we have no option, or no other way, but to test all our patients with mCRPC with germline and tumor NGS testing. Ideal is tissue, but now liquid biopsies can easily make up for the lack of tissue. If you don't have the tissue, ctDNA testing is a great platform, if you will.
They are able to identify these mutations in the HRR pathway, or TMB-high, MSI-high status, and many other tumor-agnostic approvals are based on other mutations. So, I think it's time now that we cannot say we don't have time to test our patients, because they have serious implications on treatment of our patients. Would you agree?
Pedro Barata: Absolutely, Neeraj. It is absolutely important. It is mandatory to offer both germline and somatic testing. They are complementary. One does not replace the other. We do have prognostic value. We know we can predict what's going to happen to that patient, because of what they harbor. And also we have predictive value; we have therapeutic implications based on those results.
We can also screen the family through cascade testing for the germline-positive results. So, that's one piece. The other is tissue. It is considered the gold standard if available, but it's okay if it's not available, because we can do liquid biopsy.
Regarding liquid biopsy, the timing of the testing is important. You want to test when the patient is progressing, and not after starting therapy three weeks ago, four weeks ago, because you want to find circulating cell-free DNA in the blood when you are actually doing the collection.
The future is bringing more and more biomarker-based approaches. That means more and more, we'll be dealing with what your NGS information looks like, in addition to the clinical characteristics and patient characteristics, for us to best make the best decision for our patients.
Neeraj Agarwal: Great points. Well, thank you for sharing your wisdom today, and thank you for being here.
Pedro Barata: Thank you so much, Neeraj. It's always a pleasure. Thank you.