Charles Ryan: Hello from ASCO 2019. I’m delighted today to be joined by Wassim Abida, who’s at Memorial Sloan Kettering Cancer Center where he’s a medical oncologist in the genitourinary oncology service. Welcome.
Wassim Abida: Thanks.
Charles Ryan: You’re doing a lot of great work.
Wassim Abida: Thank you.
Charles Ryan: And in particular in the area of genetics and identifying molecular genetic subtypes of prostate cancer in particular for selecting people for therapies. Tell us about the work you’re doing in particular right now around cell-free DNA and identification of the different genotypes.
Wassim Abida: So, the main goal in general that I think of is exactly what you said, which is subtyping prostate cancer so we can target therapies for patients. So, the one clear indication for such molecular testing at this point is essentially the identification of MSI-high prostate cancer because we have pembrolizumab approved. Now that I would say that that is the only indication that we currently have in prostate cancer for a standard drug to be used based on molecular subtype. But, we see other indications that are emerging.
One of those is the use of PARP inhibitors for at least BRCA-altered prostate cancer. So, it’s still considered investigational, but it does have breakthrough therapy approval from the FDA. So, for a couple of PARP inhibitors, rucaparib, olaparib, and so we do expect based on the responses that we’re seeing that that’s going to be standard of care.
So there is a big question, how do we identify these patients and how can we facilitate the identification of the patients so that we don’t miss targeted therapies for patients who might benefit?
Charles Ryan: And there are a number of ways a physician can test the tumor-
Wassim Abida: Absolutely.
Charles Ryan: … or the patient, if you will.
Wassim Abida: That’s right.
Charles Ryan: For one of these mutations.
Wassim Abida: Yeah.
Charles Ryan: Walk us through the thought process and how they do it.
Wassim Abida: Yeah, it’s actually very complex. And I would say there’s not a clear understanding across the community for how we do that because there are a lot of intricacies to it. So, number one is potentially alterations that occur in the germline. So that’s probably the simplest one. So, we know based on recent studies that essentially some percentage of men that’s between 15 – 20% will have, with metastatic prostate cancer, will have an inherited mutation.
Most typically in a gene such as BRCA and the homologous recombination repair pathway, less commonly in Lynch Syndrome type mutation. All of these are potentially actionable at this point, either PARP inhibitor or MSI high. I would say that that… And it’s become guideline actually to at least discuss doing inherited DNA testing with men with metastatic prostate cancer. That’s probably the easiest test to get.
Charles Ryan: Yeah. So most of the guidelines are actually, I think, evolving to recommend testing for patients who present with high grade or metastatic disease.
Wassim Abida: Or metastatic. Certainly the metastatic patients, and this has treatment implications, it has implications for family members. And so that is one way of identifying those patients who carry an inherited mutation. Now that’s of course a subset of those patients. But with appropriate counseling, the tests are out there. I mean, there’s several commercial companies that offer it. Certainly genetic counselors can be employed. So I would say that’s the easiest, most accessible because it doesn’t require having tumor material.
Charles Ryan: Right.
Wassim Abida: Now, of course, that only captures a portion of the men who have alterations because it turns out if you think of alterations in BRCA, half the time they happen as inherited mutations where they can be identified this way, but half the time they’re only somatic, they’re only present in the tumor. So what kind of tumor testing can we do to identify those patients?
Most of the data that we have is really from doing actual tumor testing, either from the initial biopsy material or for newly acquired tissue. And there’s a big question of what is the value of getting a new biopsy when you’ve got archival tissue? And it’s a complicated question to answer because it’s a new procedure with its own risks, certainly in men who have bone only disease, it’s trickier because we get higher failure rate from doing biopsies with tumor testing.
And so, but that is the most established method and that’s the one that we’re most comfortable with. That’s the one that the frequencies of BRCA alterations are based on. It’s basically biopsy of metastatic material. So my impression overall at this point is whenever possible a new biopsy is preferable until we gather more data.
If archival tissue is available, I think that’s the next, that’s certainly a reasonable way to get tissue. But depending on how old the archival tissue is, that may be an issue. So certainly we know that older archival tissue, initial biopsy material from five years plus has a higher failure rate as well.
Charles Ryan: So, one thing that I get a lot of questions from clinicians on, and sometimes I find confusing myself is this issue of monoallelic versus biallelic loss. So, you send the tumor away, you get your report back and you see BRCA2 alteration, right? Or whatever, some other alteration. How do you tell from that report that this is a pathogenic mutation? Now some reports will say this is a pathogenic mutation, right?
Wassim Abida: Right.
Charles Ryan: And that implies somebody’s looked at it and they understand it, but how can… What are the factors that would lead us to know that?
Wassim Abida: Yeah. So first whether the mutation is pathogenic or considered deleterious or not, that’s based on our current understanding of what effect that mutation has on the function of the protein. And that doesn’t address the question of whether you have complete loss of the protein in the tumor cell. But if we start with that, that’s based on what we understand about biology.
So if you see a mutation that is a truncating mutation, cuts off a portion of the protein where you expect the protein to lose function, at least the protein that’s produced by that allele, then that’s considered deleterious.
Charles Ryan: So not all mutations in a gene are going to be deleterious. But when you see like a BRCA2 commonly altered mutation, and when you see BRCA2, a protein with… a gene with a commonly, that is commonly mutated and you see a mutation in that you should infer or imply that that’s a pathogenic mutation because it’s a commonly mutated gene. Right? Understand the logic?
Wassim Abida: Yeah. But not necessarily because BRCA and ATM and other actually genes in that pathway are so large that missense mutations can occur as just random passenger mutations. And so, the strategy currently used to actually enroll patients on PARP studies does require some understanding of the function of that mutation. Because of the high likelihood when you see a missense mutation that hasn’t been characterized before, that it would be a passenger. Meaning that it’s not biologically meaningful and not an oncogenic or pathogenic mutation.
So, I mean, I agree with the enrollment strategy currently for these studies, which is to only allow mutations that are considered to be pathogenic-
Charles Ryan: Known pathogenic mutations.
Wassim Abida: … to the best of our understanding. So either if it’s inherited, that would be one that was characterized as pathogenic or likely pathogenic based on the genetics criteria and not the variants of unknown significance where they could turn out to be, but they’re better off to be excluded until we gather more data on it. Because you don’t want to dilute also, I mean, we’ve seen that experience happen by diluting your selected population for patients who frankly would likely, are more likely than not, not respond to the agent.
Charles Ryan: And then the issue of monoallelic versus biallelic?
Wassim Abida: Even a tougher question. And then you’re asking great questions because I get, I answer those questions from oncologists all the time. These are very complicated issues. Monoallelic and biallelic are still considered investigational at this point. The commercial tests don’t typically make that call in the CLIA setting. Certainly there is a way to identify biallelic loss on the tumor, and that may help us actually refine the biomarkers.
The biomarker currently is, do you have a pathogenic mutation in BRCA, whether it’s inherited, whether it’s somatic. However, you know, certainly we do need to refine that biomarker to really narrow down the patients who are most likely to respond. That may be part of that refinement, but it is sometimes difficult to call whether a mutation is monoallelic or biallelic based on panel sequencing. It’s more reliable if you do whole exome sequencing, which is frankly not feasible currently on a larger screening scale.
But, it’s certainly, I mean, we’re starting to gather data. And, in fact, in the data that we’re presenting today in the poster, we do make an attempt to call zygosity, meaning monoallelic, biallelic loss in BRCA.
And we find that roughly in the cases where we are able to make that call, we identified biallelic loss in 80% of the cases. So in prostate cancer at least, and this is consistent with data that we’re finding in general, when you see a pathogenic mutation in prostate cancer, more likely than not, it will be biallelic loss.
Now, in the 20% of cases where we call monoallelic loss, it’s not necessarily monoallelic. The method has limitations. You could often have loss of the other alleles through certain events that you can’t capture on panel sequencing. So it’s at least 80%, I would argue.
And so from that perspective, I think when you’re screening for a PARP study, you can reasonably assume that if you just identify one pathogenic mutation, more likely than not that patient’s going to be biallelic. So it’s a reasonable screening strategy.
Charles Ryan: So, if a clinician is out there and they don’t have tissue and they can’t get tissue and they want to do cell-free DNA tests, there are commercials cell-free DNA tests. What do those do for us and what are the limitations of them?
Wassim Abida: So there are limitations with getting tissue we’ve talked a little bit about that bone only disease, old archival tissue. And so we have to do what we can to identify as many patients as possible who could benefit from these studies. One way is doing germline sequencing. But that will miss the tumor-specific, the somatic alterations, so cell Free DNA.
Certainly, I would say it’s still early days. We’re learning more about the use of cell-free DNA to screen patients in prostate cancer. Certainly in lung cancer there’s a lot more data that already supports its use for selecting targeted therapy, GFR-targeted therapy being among that.
In prostate cancer, we’re starting to learn about how to use it. There was a very interesting paper from Ken Pienta comparing cell-free DNA findings, it was published in JAMA Oncology a couple of years ago. That was actually concerning in terms of the concordance between those methods.
So what we did in the poster here is we described our screening findings for TRITON2, which is a rucaparib study, PARP-inhibitor study for BRCA altered prostate cancer where we compare the findings of cell-free DNA to tumor testing. And essentially all the patients in this advanced patient population had tumor testing and/or cell-free DNA testing. And we find actually the identification rate of BRCA mutations to be greater than 90% in patients in whom we could obtain cell-free DNA.
Charles Ryan: Where you have a known tissue diagnosis, the cell-free reveals it in 90%.
Wassim Abida: Greater than 90%, and this is a selective population in the sense that these are all known BRCA altered patients who went on the study.
Charles Ryan: Right.
Wassim Abida: And so we essentially, in those patients we know they have a BRCA mutation, how often in this screening set were we able to find a mutation in BRCA in cell-free DNA greater than 90% of the cases? Now of course, some of those cases are actually germline events where you would certainly identify the BRCA mutation. But in half of those cases it’s a non-germline, it’s a somatic event.
So, we’re starting to get information that cell-free DNA can be reliable to identify patients. It will miss deletions in BRCA.
Charles Ryan: Yeah.
Wassim Abida: So that’s roughly 20% of cases. At this point you cannot call deletions and you will miss those. And this is, I just want to say that this is an advanced prostate cancer patient population. And we know that cell-free DNA essentially sensitivity is dependent on disease volumes. So in earlier disease states we may not get that level of sensitivity.
Charles Ryan: Right. Great. Well, this is really interesting. I mean, it sounds like there’s progress being made in this area despite Ken Pienta’s warnings a couple of a couple of years ago. We’re getting better, smarter-
Wassim Abida: Yes, I think so.
Charles Ryan: … at analyzing this and the commercial assays are perhaps coming along. And I think the great thing is we’re also doing this in the context of the development of a new class of drugs, the PARP inhibitors.
Wassim Abida: Right. Absolutely.
Charles Ryan: So, congratulations on your achievements and your work on that front-
Wassim Abida: Thank you.
Charles Ryan: … and for your insights today. Thank you very much.
Wassim Abida: Thanks again, Chuck.