Genetically Informed Treatment for Advanced and Metastatic Prostate Cancer - Alicia Morgans
April 6, 2020
Alicia Morgans, MD, MPH Associate Professor of Medicine in the Division of Hematology/Oncology at the Northwestern University Feinberg School of Medicine in Chicago, Illinois.
The Biology of Prostate Cancer - Kenneth Pienta
Article: Germline mutations in prostate cancer carcinogenesis
Written Conference Coverage: Disparities in Prostate Cancer Management in the United States
Alicia Morgans: Hi, my name is Alicia Morgans and I'm an Associate Professor of Medicine at Northwestern University. I am going to talk today about genetically informed treatment for advanced and metastatic prostate cancer. Really an area of a lot of progress and change both clinically, in terms of therapeutics but also in terms of the implications for our patients and their families. So let's begin.
So a lot of people ask really what are the most important germline and somatic mutations that we should be thinking about for our patients with metastatic prostate cancer. And I think to start, we should acknowledge that we should be thinking about both germline and somatic changes. The changes that occur in the germline are things that are going to be passed down from our patients to their children and from their parents to them. And things that are somatic are going to be mutations that occur only in the tumor and have to be tested by either using circulating tumor DNA or by getting a tissue biopsy and assessing them that way.
So it's important for us to acknowledge that both of these are going to be important components of the care paradigm for our patients at this point, particularly as we expect to have some potential approvals for things like PARP inhibitors in the coming months that will then be very closely linked with therapeutic options for patients with some of these mutations. All patients with metastatic prostate cancer at this point should have germline testing too. Again, these are the inherited mutations that patients have gotten from their parents and can pass onto their children that may also be present in any siblings that our patients have. So this is important and recommended for all patients with metastatic prostate cancer at this point.
So when I think about which mutations are going to be most important to test for our patients in both germline and somatic settings, I think about the ones that are listed here on the screen. Most importantly is going to be BRCA2, as it's got both the highest prevalence in the germline, estimated here at about 5%, 5.5%, as well as the highest prevalence for the most part in the tumor itself around 5% as well. But we also have to remember that there's also BRCA1, which is also present about 1% and the germline and in the tumor tissue. PALB2 which, it appears to be more common in the tumor than in the germline but is possible certainly in either one. ATM, also present, you can see here maybe a little more common in the tumor. And then MSH2/6 which actually have implications for microsatellite instability and for use of things like pembrolizumab or other checkpoint inhibitors. So pembrolizumab at this point is the only one with an indication for all patients with microsatellite instability, which is why I've mentioned that specifically.
But, the key message here is that germline testing is recommended for all patients with metastatic prostate cancer. Germline and somatic testing can both actually have implications for patients in terms of their therapeutics and the main mutations that I think need to be assessed if nothing else for our patients are the BRCA1 and two genes, PALB2, ATM, and MSH2/6 which has implications for microsatellite instability.
Our genetic counselor at Northwestern and I have recently published a paper thinking about priority genes of interest and how we can think about incorporating testing into our clinical practices and I'll mention some more details of that later on in this talk. But this is just an example of the many genes, and this is only part of the table that we can test for and the implications in terms of therapeutics with PARP being PARP inhibitors and CP being checkpoint inhibitors, particularly pembrolizumab, as that has the indication for any patient with microsatellite instability.
So many genes for us to think about and if we are able to do a more comprehensive panel for germline as well as next-generation sequencing for sematic testing, we can think about many more mutations that may have relevance in addition to those main ones that I mentioned before. But if we do find a mutation in our patients in DNA repair defect or in microsatellite instability genes, are there implications for sequencing with the therapies that we currently have available to us for prostate cancer and do these germline test results actually change my practice as we stand today in this current landscape?
Well, let's look at some of the data. There was a recently published really interesting paper called PROREPAIR-B. It was a study that was done in Spain led by Elena Castro and her group and really fascinating look sort of at a couple of 100 patients and monitored them in a non-randomized way just as they were receiving standard of care therapies for metastatic CRPC. And these patients were treated with taxanes and with AR targeted agents and Dr. Castro and her team looked at the cause-specific survival in these patients.
On the left, we can see the cause-specific survival curves for patients who had ATM, BRCA1 and BRCA2 in the red line. And the other line, this was a non-carrier population and you can see that the team and this was the primary endpoint of the paper, the team lumped together all patients with the most the DNA repair defects of most interest at the time that this study was done and that was ATM, BRCA1 and BRCA2. What we can see in that cause-specific survival curve on the left as they followed these patients as they received standard treatment. Again, non-randomized study, they could receive whichever treatment was indicated and prescribed by their physician. There was actually not a statistically significant difference in cause-specific survival, meaning prostate cancer survival, in patients who had these DNA repair defects versus patients who did not.
But interestingly, when Dr. Castro and her team pulled out specifically the patients with BRCA2, which again is the most common germline mutation that we find in one of the most common somatic mutations that we can find as well. When they pulled those patients out and they compared them to non-carriers, there was a statistically significant difference with the BRCA2 carriers in the teal-colored line having a poor cause-specific or prostate cancer-specific survival when compared to non-carriers in the blue line.
So important that if we look at all-comers with these DNA repair defects, there may not be a difference, but that BRCA2 may have a particularly poor survival in this setting. So something for us to think about and that's what makes this question of whether these DNA repair defect mutations actually affect our practice. That's what it makes it so important that there may be some poor survival in patients who have BRCA2 mutations in particular.
So interestingly in this study, Dr. Castro and her team looked then specifically sort of an exploratory analysis to try to think about whether patients who had BRCA2 did better or worse if they use different sequences of therapies. So to go through this on the left, again we're looking at cause-specific survival of BRCA2 carriers in the red line and patients who did not have DNA repair defect mutations in the blue line. And what Dr. Castro's team looked at in this graph on the left, or figure on the left, is if patients were treated with an AR targeted therapy followed by a taxane, was there difference between their survival and those patients who had no DNA repair defect mutation. And you could see these lines completely overlap when patients received an AR targeted agent followed by a taxane. And again this is non-randomized, so patients were receiving the treatments per the discretion of their physicians at the time.
So if we look on the right, again non-carriers are in blue and BRCA2 carriers are in red. We're looking again at the cause-specific survival. But in this setting, patients are all receiving a taxane agent, docetaxel in this case, followed by an AR targeted agent. And we can see that these curves actually split pretty dramatically suggesting a poorer survival. And this is a statistically significantly poor survival for patients who are treated with a taxane followed by an AR targeted agent if they had BRCA2 mutation as compared to non-carriers. So very, very thought-provoking, interesting difference here with a poor survival being in those patients who had the taxane first followed by the AR targeted agent when they had BRCA2.
I should mention that the patients who were treated with the taxane first actually had poor disease features and risk features. So they had more aggressive appearing disease than those patients who had an AR targeted therapy first followed by a taxane. Remember this is a non-randomized trial, so there were definitely differences in patients who may have gotten a taxane first versus those patients who had an AR targeted agent first. And so this is really thought-provoking, but from my view should not change our practice because it's a non-randomized trial, and it's really more of an exploratory analysis, but raises the question of whether sequencing really may matter in patients who have BRCA2 mutations.
Interestingly, Emmanuel Antonarakis and his group at Hopkins also did a study looking at patients who had any DNA repair defects versus non-carriers and assessed them in standard of care assessment, not a randomized fashion, but assessed them in terms of their survival when they were treated with an AR directed therapy. And what they found, really fascinating, was that patients who had a DNA repair defect mutation with the blue line actually had better survival when treated with an AR targeted agent than patients who were non-carriers. So again, I don't think given the number of patients, and you can see the patients at risk here at the bottom being somewhere around 175 to start, I don't think that this relatively small study should make us always give AR targeted agents to patients who have DNA repair defects as a first-line therapy. But it raises the question of whether this may be the case. This was not a statistically significant difference though I should say.
So importantly, I think we're all aware of a landmark trial that was recently completed, the PROfound study, which was a Phase III trial that selected four patients who had DNA repair defects. They had two cohorts, these patients had all received treatment in the past with an AR targeted agent and most of them, more than half, had also received at least one chemotherapeutic agent as well. Some of them had had both, but all of these patients had had progressive disease on abiraterone or enzalutamide and many had also progressed on docetaxel. They were selected if they had BRCA1, 2, or ATM mutations for cohort A and if they had one of a number of other DNA repair defect mutations, they could be in cohort B. Patients who were not carriers of any of these mutations were actually not allowed to be in this trial.
The two cohorts were randomized in a two to one fashion to receive the other AR targeted agent, so if they had progressed on enzalutamide they received abiraterone, or if they had progressed on abiraterone, they received enzalutamide, or they received olaparib. And this was a two to one randomization. So two out of three patients had olaparib and one out of three patients had the other AR targeted agent.
They were followed for radiographic progression-free survival, and then there were actually crossed over if they wished if they were given the AR targeted agent. They crossed over to olaparib after progression. And you can see here this is the schema outlining that trial.
The primary endpoint was radiographic progression-free survival in cohort A alone. So those were patients who had BRCA1, BRCA2 or ATM mutations. And what you can see here is this dramatic separation of curves. We could see the six month survival as well as the 12 month survival in landmark analysis here. But this beautiful separation of curves demonstrating superior radiographic progression-free survival for patients treated with olaparib in the red line as compared to those patients receiving AR targeted agent in the blue line and really a dramatic separation in these curves.
Interestingly, they also had, was not statistically significant, but it appeared that they had improved. They will probably have improved overall survival as well, but again that did not reach statistical significance and there was at least a 75-80% crossover rate in that analysis as well. So they maintained what appears will be a statistically significant survival benefit even with a dramatic amount of crossover from the AR targeted agent on the time of progression to olaparib.
So what I would say besides the fact that olaparib looks to be a very promising agent associated with improved radiographic progression free survival in patients with DNA repair defects who had these BRCA1, BRCA2 and ATM mutations, we definitely have more work to do. We have to understand better how different mutations may affect radiographic progression-free survival and survival itself.
And what I have on the screen now are illustrations of some of the exploratory analyses that have been done in the PROfound study on the right and the TOPARP-B study on the left, which was another study selected for patients with BRCA1, 2 and other DNA repair defect mutations who were treated with olaparib. What we can see on the left is that in the TOPARP-B study and these patients were treated with olaparib, there was a difference in their composite response rate whether they had BRCA2 and 2 mutation and ATM mutation, a CDK12 mutation, PALB2 mutation and others. So important that we can see different responses in terms of resist response. PSA decline greater than 50% and CTC conversion depending on the different mutations. With ATM, for example, having more higher rates of CTC conversion but not really resist responses, whereas BRCA1 and 2 appeared to have a pretty high rate of resist response, so shrinkage of measurable soft tissue disease and high rates of PSA responses and CTC conversions, which is just different than ATM and CDK12, of course. PALB2 interestingly in this study may suggest that there's some response similar to BRCA1 and 2, so PALB2 though somewhat rare appears to have some nice responses and this setting to olaparib.
When we look on the right, this is an exploratory analysis from the PROfound study that we just mentioned and in this we can see again when patients were treated with olaparib in the purple-ish red bar, there was actually a much greater response than physician's choice, which is the AR targeted agent for patients with BRCA2 mutation and CDK12 mutation but perhaps less so for BRCA1 and maybe ATM.
So this is all exploratory and something that I think we need to continue to look into and these numbers are actually quite small for most of these mutations but really fascinating for us to think about and important implications to come in our clinic though they may not be here today.
What it would also say is that patients with DNA repair defects also appear to have some response in some settings to platinum though these are non-randomized trials, these are not prospective studies and we don't actually have a large prospective study of platinum in a selected population yet to really guide our care. But we can use platinum in these settings for some patients with at least some indication that this may be a benefit to them. On the left, we can see Mark Pomerantz and his group at the Dana-Farber's collection of patients who have had BRCA2 mutations in the orange lines versus non-carriers in the blue lines and these patients are treated with platinum and we can see that some of them seem to have a pretty PROfound response.
What we could see also is that many of these patients actually had a PSA decline whether they were carriers or not carriers. So it raises questions about how we're selecting these patients, but in any event, platinum may show promise and on the right, a really interesting three patient study, which is really just a sort of a case series of patients who had biallelic inactivation of BRCA2. This is Heather Ceang's group out at the University of Washington and they just show some pretty dramatic responses in terms of PSA for these patients when they were treated with carboplatin. So platinum shows promise, but certainly is not there yet for us to use. I would say routinely as one of the early lines of therapy for our patients with DNA repair defects, but maybe an option should we need something and options are running thin in terms of standard FDA approved options.
Also really important and interesting. We've had a lot of excitement around the use of anti-PD-L1, drugs or anti-PD-1 drugs. This is a particular study that Julie Graff and the group over at Oregon Health Sciences University did and Tom Beer as well. They looked at patients who had had progression on enzalutamide continued the enzalutamide and actually started them on pembrolizumab and three out of 10 patients actually had really dramatic responses to some radiographic responses and not all of these patients had microsatellite instability. And this is the indication currently for pembrolizumab of any patient with microsatellite instability.
So race questions has led to multiple studies looking at patients progressing on enzalutamide to see whether pembrolizumab may be a reasonable option and in this patient population, and we can see here there are several keynotes studies that are looking at this. This is one that was recently presented by Johann de Bono, patients with metastatic CRPC who had progression on some AR targeted agents and chemotherapies and they were started on treatment with pembrolizumab at the standard dosing for every three-week cycles. What we can see here is that although there's nothing really dramatic in terms of CRs and PRs, complete responses and partial responses, there was some stabilization of disease in patients who were in these two cohorts, these patients could have had PD-L1 upregulation or not, which is interesting and important to know.
And further study KEYNOTE-365 and this was presented by Evan Yu, a little bit more recently in 2019. These patients were treated with pembrolizumab and this was combined with olaparib and this was again patients who had had docetaxel for mCRPC and they had had progression and what we could see here is that there may be a slightly higher response rate, a little bit higher response rate and perhaps if we are able to identify patients who have DNA repair defects, we may see that this combination of olaparib plus pembrolizumab may be able to be even more effective in them.
What I think is interesting here is that there were responses again in patients who were PD-L1 positive and there were more responses in that population than those in the PD-L1 negative population. But there were some good PSA responses at least even in patients who had unknown or low expression of PD-L1, so work to be done. Nothing earth-shattering at this point, but this is definitely an area of interest.
So do germline test results change my practice if I find a DNA repair defect? So they don't really strongly affect my choice of agent in metastatic hormone-sensitive prostate cancer. Maybe I would be less likely to use chemotherapy as a first-line treatment option based on the PROREPAIR-B study for patients with metastatic CRPC. But this study was, again, not randomized, numbers were small and I think that it's more hypothesis-generating than anything else. And so really still I have conversations with my patients about making sure we switch mechanism of action when we're going from metastatic hormone-sensitive to metastatic CRPC and talking to patients about their preferences for side effect profiles and ways of, all of the things that are different about all of these agents. But I wouldn't necessarily choose an AR targeted agent over chemotherapy, I should say based on that study. But definitely it's something to think about.
We can, I think, use olaparib. At this point, it would be off-label. So that's going to be kind of a comfort level and something that we can think about using olaparib potentially in a post-abiraterone or post-enzalutamide treated patient population for a metastatic CRPC rather than a second AR targeted agent. I think that the radiographic progression-free survival that we saw on that Phase III study suggested that there's a clear benefit to using these targeted agents over another AR targeted agent. So important for us to think about and I do expect that we may see that olaparib may be approved at some point in the future and we can use it on label at that point, but something that we should be on the lookout for this to be coming down the line.
And platinum may be an option for DNA repair defect patients, but I would say this is a late option, not something I would move up into early, early lines of therapy because we just don't have the prospective data to demonstrate the clear benefit there. For patients who have microsatellite instability and evidence of mismatch repair mutations, I think that pembrolizumab is an option. We know it is approved for patients with microsatellite instability. Something for us to think about. I usually think about this more as a clinical trial option or a later line option for patients with metastatic CRPC. But always something to have, at least in our thoughts. If we identify patients with this, and this is only going to be about 3% of patients, or so we believe, with metastatic CRPC. So not a large population, but this can be an important option for those patients.
What about challenges to implementation? I think that incorporating germline and somatic testing into our clinical workflow is not something that has been easy for our practice and is not necessarily incorporated into all practices. So what do we need to think about? So our group in the Prostate Cancer Clinical Trials Consortium actually put together a survey to understand what some of the challenges and barriers to implementation of germline genetic testing might be in practices. And this was led by Channing Paller at Johns Hopkins as well as Heather Cheng at University of Washington.
And of these 19 institutions who responded, 26 oncologists reported, there were five primary barriers to implementation and these are in institutions where there was actually sometimes a lot of implementation or a lot of integration of germline genetic testing into clinics already and some institutions that were not quite there yet. So an interesting slice of the academic pie where these things were more or less common, but the things that were identified were a lack of access to genetic counselors for either pre or post-test counseling, concerns about insurance coverage and out of pocket costs for patients, clinic workflow challenges. When do we do the testing and how do we get it to work effectively in our clinical workflow, time and space availability for counseling before or after testing, and resource availability for clinicians and patients to understand the implications of the testing that was happening.
In our system, we have been so fortunate to have a genetic counselor that was identified and hired through our urology department and so there may be other ways that people are going to do this, but in our institution, we actually have funding through the Urology Department as we tried to make into implementation of these recommendations to do germline testing financially viable in our system. And we are also partnering with our mid-level providers to try to enhance billing for the counseling that's happening because that can be a challenge.
But in any event, for institutions that do not have a genetic counselor and importantly many of the germline genetic testing companies actually have online chat boxes and other methods to engage patients and clinicians online and in most cases for free to go through post-test counseling results for patients with positive testing. So there are ways to try to do this without into implementing this in the clinic. But if you can include a genetic counselor, I think that it is so, so helpful to the clinicians and to the patients if that's possible.
We are doing all of our consenting for genetic counseling currently ourselves as oncologists and urologists are actually doing that, too. And our genetic counselor is following up positive results and then counseling, post-test counseling, only for positive results and we're doing our own pretest counseling as clinicians, as oncologists and urologists. But if patients have a very strong family history or they're particularly challenging, we do engage with our genetic counselor upfront for pretest counseling as well. And then as I said, there can be phone counseling if there are just questions or if there are variants of uncertain significance, which I think tend to lead to also a lot of questions and stress sometimes for patients. These are not things that we act on clinically, but sometimes it can be concerning for patients if they see there's a BRCA2, a variant of uncertain significance, for example. So it's nice to have either a genetic counselor or to use these chat boxes with the companies that do the testing.
And importantly to relate it to the question about insurance coverage, Invitae is one of the companies that does this testing and they have been providing free testing now for patients who meet their financial approval metrics and so patients can fill that out and I would imagine that other companies are going to be doing that as well. This is just one that was available or is available currently.
So in conclusion, implementation of novel scientific advances is challenging and requires time and investigation of best practices. And this is what we have here. This is germline and somatic genetic testing for patients with prostate cancer is an evolving area of importance in our clinical workflow and we have to figure out how to best do this and people are actively investigating this.
Priority genes to test include these HRR genes or the DNA repair defect genes and microsatellite instability, mismatch repair deficiency mutations, as well as there are implications for PARP inhibitors, potentially platinum agents and for a microsatellite instability for pembrolizumab as well.
Treatment implications really remain a work in progress and there are prospective studies that are going to be critical for PARP inhibitors, olaparib is currently being reviewed, may be approved by the FDA at some point in the future based on PROfound or other data. And platinum I would say is still being studied and hopefully there will be prospective data at some point so that we can potentially use that in these patients as well.
Metastatic hormone-sensitive disease is really primed for study, is an area where I think some of these treatments may be even more effective just as we've seen with other therapies, chemotherapy and AR directed therapies as we've moved them forward. So an area for us to keep our eye out for as new studies are coming down the pike. And workflows that incorporate post-test counseling with a genetic counselor or maybe even a chatbot, may be the best use of resources as we all try to figure out how to incorporate germline and somatic genetic testing in our clinical practices. Thank you so much for your time and attention today.