Read the Full Video Transcript

Charles Ryan: Hello and welcome. I am delighted to be joined today by Wassim Abida from Memorial Sloan Kettering, where he is an assistant member of MSKCC and an Assistant Professor of Medicine at Weill Cornell Medical Center, Weill Cornell University, and a medical oncologist, and he worked on a really interesting abstract that was presented at GU ASCO that looked at the co-occurring gene alterations and rucaparib in patients with BRCA1 or BRCA2 mutated metastatic CRPC. Wassim welcome. And tell us a little bit about the work that was done here.

Wassim Abida: Hi Chuck.  Thanks for having me. So yes, the abstract was related to really trying to understand what are the modifiers that impact response to the PARP inhibitor rucaparib from this prospective study. And I just broadly want to summarize where we are. Luckily now we have two PARP inhibitors approved in metastatic CRPC. Rucaparib and olaparib based on prospective phase two and phase three studies. The indications are a little bit different. As we all know, rucaparib is approved for BRCA2 and BRCA1 mutated metastatic CRPC and olaparib is approved for a broader set of gene alterations, which at this point, we are starting to understand that they absolutely are not all the same. That BRCA2 in particular really has the strongest association with the response, but some of the other genes have fairly low response rates to really no data. And yet the PARP inhibitor olaparib is approved for them.

So I think we are starting to scratch the surface of really what specific genomic features predict for a response, and we really need to improve on that biomarker. And our goal here was to look at the impact of three frequently altered tumor suppressor genes on the response in the BRCA2 and BRCA1 settings. And so we looked at p53 mutations, which occur in about 50% of mCRPC, PTEN mutations which occur in about 30% to 40%. And RB1 mutations are mostly lost, which occurs in about 13% to 15% of mCRPC. And we were interested in seeing if alterations, one of these tumor suppressor genes, which has been associated with worse outcomes in general, can impact response to PARP inhibitors.

Charles Ryan: So what is your take-home message for clinicians who are out there today seeing a CRPC patient and they get a report back and it indicates if there is a BRCA alteration. But then there are all these other alterations. How should they use that in their practice in terms of, in which cases, should they say a PARP inhibitor is appropriate, or are there cases where they can look at other alterations and say, a PARP inhibitor may not be appropriate?

Wassim Abida: My belief is absolutely. So we know that the FDA approved in particular, olaparib for a broader set of genes, but in my practice, the gene is very, very important in all of this. So BRCA2 alterations and even a BRCA1 alteration, I would recommend a PARP inhibitor. In particular with BRCA2, I would be specific to the patient that the likelihood of benefiting is high, and it's about 50% of having an objective response, some additional percentage with stable disease for some period of time. Now, if I have a patient with an alteration in ATM or CHEK2 CDK12, then I, it's not that I would exclude trying a PARP inhibitor, but certainly, I would look for alternatives.

I would tell the patient that although the PARP inhibitors are approved in that setting, that response rates were extremely low. And that, that has been now born out by multiple studies. And certainly, I have put patients particularly at, while the studies were ongoing with these alterations and I can't say that we really are seeing a significant benefit. And so, although there is an approval there, I think that just like all the discussions that we have as oncologists, we have to temper expectations based on what we understand about these genes.

Charles Ryan: Yeah. And you've used a couple of keywords there. So first of all, before we get into that, BRCA2 pathogenic variation, pathogenic alteration is a predictive biomarker for response to a PARP inhibitor. So, that would be your reason to consider a PARP inhibitor. But the key term that you used is a moderating factor or a tempering factor. And that may be the RB loss of the TP53 alteration. And that is, I think a key message for clinicians.

Wassim Abida: Right. So as far as moderating factors, if we are talking about BRCA2 specifically, I can't say right now that really any factor should necessarily impact our decision to use the PARP inhibitor. Now in our own study, and mind you, we are looking at the study of 115 patients evaluable for PSA, 62 evaluable for radiographic response. So, not a tiny dataset, not that being said, we are not talking about hundreds to thousands of patients. We did not see a clear association for loss of p53 or PTEN with a response in then the BRCA mutated setting. So, my conclusion here is that mutation p53 or PTEN really does not appear to impact response to PARP inhibition for BRCA two and BRCA1 mutated metastatic CRPC. Now RB1 is, we have fewer cases and that is the caveat, but there does seem to be a signal that it may impact response to PARP inhibitors.

And I think we just need bigger data sets and, and we will get there. We just need to aggregate data sets. And now with RB1, RB1 sits right next to BRCA2 chromosome 13. So, frequently, you can get loss of both genes co-occurring and we know that RB1 loss is associated with neuroendocrine phenotype as well. So is it possible that loss of RB1 or the neuroendocrine phenotype may dampen the response of PARP inhibitors? It's possible, I think we need to investigate this further. I would say we are scratching the surface of identifying really modifiers of response. And I think we certainly, it's a big development in terms of the approval of PARP inhibitors for BRCA mutated and other mutated mCRPC. But I think we definitely need to refine the bar marker, and we need more studies like this that are bigger.

Charles Ryan: So it's a really important point. And it's, I want to come back to RB, but it's a really important point about the distinction between what might be a prognostic factor and a predictive factor for the clinicians out there because I mean, we pretty much know that a TP53 alteration is a poor prognostic factor in prostate cancer. But what you just said is it's not necessarily a predictive factor for an adverse outcome on a PARP inhibitor.

Wassim Abida: That's correct. That's correct. There was not a significant difference in responses to rucaparib in the BRCA mutated setting for patients who are p53 mutated. And we have reasonable numbers for p53 mutation because that occurs in about half the cases.

Charles Ryan: Right. And that is a key point. So I doubt there are that many clinicians out there saying, well, I would normally give you a PARP inhibitor, but because you have a TP53 alteration, I'm not going to.  But I think that we can think about the response in the short run and prognosis in the longer run. And as oncologists, we need to be disciplined about tempering our enthusiasm over the long haul in these complex cases. I want to circle back. RB1 is interesting and its co-location with BRCA2 is very, very interesting. And so do you have a sense of what proportion of BRCA2 altered patients, I mean, you have this in the abstract, BRCA2 altered patients also have an RB1 alteration, I think it's around 15% or so.  But going the other way, if we were to take a population of RB1 loss patients, how many of them would have features that would potentially include a BRCA2 alteration, and, or I'd like you to comment also on LOH and the relationship to PARP inhibitors if you can.

Wassim Abida: Sure. So, I'll answer the first one, the first question, which is what percentage of patients with RB1 loss would have BRCA alteration? I would say a relatively small fraction. So there is not a clear link for a complete loss. And then I think it's relevant to your second question of loss of heterozygosity. So it's frequent that you have loss of one copy of both genes because they are so closely located. So you could have, for example, loss of one copy of RB1, loss of one copy of BRCA2, and what we call shallow deletions like this or heterozygous loss is particularly common in mCRPC, which has a highly disorganized genome. You have a lot of copy number alterations. But to have a complete loss of RB1 and BRCA2 co-occurring is an infrequent event. Now, the question is, does having heterozygous or incomplete loss of both genes, does that potentially sensitize the PARP inhibitors and vice versa?

We don't know exactly. There was an interesting study by Goutam Chakraborty and published in Clinical Cancer Research that suggests that having shallow loss of BRCA and RB1 may sensitize to PARP inhibitors actually sensitized, but that has not been born out. So for now, what we are working with is having a full deletion of BRCA2. And that is what is required to actually put somebody on a PARP inhibitor. So, there is a difference there. And I should add to that, that really most commercial sequencing assays do not report the shallow loss or heterozygous loss, they only report complete deletion. So if you see BRCA2 deletion reported on, say a Foundation report or a Tempus report, it typically represents deep deletion, and then the patient should get a PARP inhibitor absolutely. But it's unusual to have both RB1 and BRCA2 complete deletion at the same time.

Charles Ryan: So what are your thoughts on the genomic reports we get from commercial sources that will say, genomic loss of heterozygosity exists. And then on the right side of the form, it will say, consider a PARP inhibitor. I've seen this over and over again. And I'm just wondering if this is accurate.

Wassim Abida: Yeah. I would say that this is just like with the initial rollout of cell-free DNA assays, there are still some technical issues to be sorted out. So calling deletions is highly dependent on the quality of the sample and the tumor purity. It's very difficult to call often if you have just one copy loss of one copy of BRCA2 in a tumor sample that may be mostly normal tissue. And so, it is difficult to call the loss of heterozygosity and these are very early days. And if we are talking about a mutation in one copy of the allele and then some evidence of loss of the other allele, then I think that that is more convincing, and that may be a better sign that the patient may respond, but these are early days for calling loss of heterozygosity. I should add that the studies so far that have led to the approval of the PARP inhibitors have not required proving loss of heterozygosity.

So you can have just one copy mutation, truncating mutation, and BRCA2, and that would be sufficient. However, we know that if you, and there was a big paper in Nature, that was pan-cancer that showed this, that if you do have, if you confirm the loss of both alleles and they did the stroke through a whole-exome sequencing, then that is a predictor of a better response to a PARP inhibitor. So when we get to a stage where we can reliably call the biallelic loss of BRCA, I think will have a better biomarker. I just don't think that we have that right now. [crosstalk 00:12:04].

Charles Ryan: And there are some of the early studies with other PARP inhibitors where they required biallelic loss and they didn't report substantially higher response proportions. I'm speaking specifically about the niraparib data if I'm remembering that correctly.

Wassim Abida: That's absolute, you're absolutely correct. I think that the issues there were number one, I think that slowed enrollment because it was sometimes difficult to confirm the biallelic loss. And secondly, you are right, the responses were not substantially different. And that is because really BRCA2 is an oncogenic gene in prostate cancer. So if you have a BRCA2 deleterious mutation, it's been shown that in most cases you have a biallelic loss. So, I think if you average it out, you may not necessarily need to look for biallelic loss, you can pretty much assume that in most cases you do, but if we are really looking to refine the biomarker and we have a reliable way to call it, then I think we will be able to refine the biomarker eventually. I would add that probably the assays probably were not refined yet in terms of calling biallelic loss. And so, that is being worked out through a computational pathway. But the quality of the samples is very, very important to do this.

Charles Ryan: So on the topic of the quality of samples, my final point I want to address is cell-free DNA versus tissue and the issue that you have published on extensively and have really been a leader in which is progressive treatment associated, acquired somatic alterations. And what do you do in your practice in a CRPC patient when you want to know what's going on genomically with them.  Are you going to their prostate sample from several years ago? Are you relying on cell-free DNA? Are you trying to get both?

Wassim Abida: Yeah, it's something that I do deal with on a daily basis and I will tell you that we know for fact that for older prostatectomy samples that we do run into quality issues. And so I usually avoid going for samples that are frankly older than three years for various reasons. Number one is I want to acquire a new sample in case there are new alterations. We actually showed that in microsatellite unstable prostate cancer, that you can get the acquired phenotype later in the disease course. So, going to an archival older sample may miss something actionable.

And that may occur for BRCA, although so far we've seen BRCA be an early event, I'm sure there are exceptions out there. So the bottom line is I look for a sample that is no older than three years old. And if that requires a new biopsy, then that is what we aim for. I still, at this point, and that may change, I still trust tissue sequencing to capture more alterations than cell-free DNA does. And that is really, the deletions are usually better captured in tissue. Cell-free DNA is developing to a point where you can call deletions, but I favor tissue. Now, if tissue is not obtainable and that does happen, then I have a reflex to cell-free DNA.

Charles Ryan: Yeah. Great point. And of course, there is some value in the histology as well. If you can get it, in particular, to rule out neuroendocrine features, et cetera. I think you have given us a really clear direction on how we can think through some of this complexity in the clinic. It is great to have PARP inhibitors available, as you said, it's early days. And I think that we are going to see, my hope is we're going to see our informed decision-making go in both directions. In other words, find cases perhaps where PARP inhibitors and other agents may be useful where we are currently not using them. And perhaps more information in the other direction, which as an oncologist, I would say more information that can help us to temper our enthusiasm appropriately when it is required to be tempered. Let's put it that way.

Wassim Abida: I agree completely. I think we, I think it's a good start what we have already, but there is definitely room to go in terms of putting more patients on PARP inhibitors who may respond and then narrowing the equality by biomarker to that. Yeah.

Charles Ryan: Yes. Yes. Well, thank you for your time today and for informing us of your work, which continues to press new boundaries for how we can interact with our patient's genomic report and the available therapies. And I look forward to hearing more from you and talking to you more in the future.

Wassim Abida: Thanks so much, Chuck. I appreciate it.