Optimal Integration of PARP Inhibitor Monotherapy into the Management of Metastatic Castration-Resistant Prostate Cancer — Johann S de Bono, MB ChB, MSc, PhD, FMedSci
PROF DE BONO: Hello. My name is Johann de Bono, and I want to talk to you today about PARP inhibition for advanced prostate cancer. I’ve been working with PARP inhibitors for really almost 2 decades.
Here’s my talk overview. I want to give you a brief introduction, then I’ll talk about the investigator trial I ran called TOPARP, then I’ll talk about a trial I designed called PROfound that led to the approval of olaparib for treating metastatic castration-resistant prostate cancer, or mCRPC, then I’ll talk briefly about some other agents, including talazoparib and rucaparib. And I’ll conclude at the end with some overall conclusions, touching briefly on the most recent data from MAGNITUDE and PROpel.
So here’s the introduction. So we have shown that prostate cancer is a highly heterogeneous disease with some intrapatient heterogeneity and much interpatient heterogeneity, with probably 1 in 5 to 1 in 3 prostate cancers having DNA repair defects that may sensitize to drugs, including PARP inhibition and platinum-based chemotherapy.
Important to know that some of the prostate cancers have other DNA repair defects, including mismatch repair defects that can sensitize to immune checkpoint inhibition and also CDK12 defects. And biallelic loss of both MMR and CDK12 can result in profound responses to immune checkpoint inhibitors like pembrolizumab.
You will also be aware that the commonest mutation in DNA repair is actually TP53, I guess the guardian of the genome, although at present we do not yet have drugs that are really focused on targeting cancers with those alterations.
It was hypothesized many years ago that DNA repair defects can actually be an Achilles heel for cancer cells, and specifically that BRCA or BRCA-associated genes, these are tumor suppressors, including BRCA1, BRCA2, PALB2, and RAD51, can have a synthetic lethal interaction with oral drugs targeting PARP, PARP1 and PARP2; PARP inhibitors that can result in tumor-specific cell kill that is sparing normal cells, really arguably the holy grail of cancer therapy.
Now I should say that ATM does also sensitize to PARP inhibition and is synthetic lethal to PARP inhibition, although somewhat less than BRCA loss. But the ATM is only indirectly involved in homologous DNA repair and is not directly involved in HRD, but actually senses DNA damage and activates the repair process, including the BRCA gene activation.
These 2 seminal papers, published almost 20 years ago in Nature, Farmer et al and Bryant et al, were the first to show that homozygous loss, that is biallelic loss, of the genes BRCA1 and 2 caused synthetic lethal interaction with PARP inhibition, that is complete loss of these genes’ function, or these proteins, resulted in sensitization to PARP inhibition. Whereas half loss, if you still have 1 allele or normal gene function, did not sensitize to PARP inhibition, and this really was the basis of the trials that I will describe from here.
Now synthetic lethality is best defined as tumor-restricted loss of both alleles, all protein function, in the tumor cell causing an acute vulnerability, or Achilles heel, of that tumor cell through an intervention that spares the normal cell that actually has a full complement or partial complement of that protein and those genes, 1 or 2 copies. And this results in selective tumor cell kill while sparing the normal cell.
Multiple groups have now shown that there are a number of key DNA repair genes that can sensitize to PARP inhibition, these being either directly or indirectly involved in homologous recombination. And the cell-killing mechanism is thought now to be PARP trapping on DNA. The PARP protein complex gets trapped in the DNA, it’s bulky, and it causes stress on that DNA, resulting in a generation of double-strand breaks in the DNA, which cannot be well repaired because of the HRD loss, the underlying BRCA loss, and therefore this results in tumor cell kill. And BRCA2 is the posterchild gene, but you can see that multiple other genes can sensitize, including, for example, ATM, XRCC2, XRCC3, and multiple other genes.
Importantly, loss of genes involved in nonhomologous end joining, for example KU70 up here at the top, can actually cause resistance to PARP inhibition, which may be an important point. Essentially though, multiple losses of DNA repair defects are found in prostate cancer can sensitize to PARP inhibition in the synthetic lethal fashion.
Now this 65-year-old patient of mine, who was a famous actor, talked very widely about his disease, is really a posterchild of this therapeutic strategy, one of the first men with metastatic CRPC with a BRCA alteration treated with a PARP inhibitor.
In fact, this man has a very interesting story. He was diagnosed with metastatic CRPC. He progressed rapidly on LHRH analogs and bicalutamide, as was the standard then, almost 20 years ago. He then actually got on trial with docetaxel with an antibody to IGF1R, progressing quite quickly after that. He got eribulin, which is now approved for breast cancer but has never been approved for prostate cancer, which he progressed rapidly through. He then got radium-223 on trial, progressed quite quickly on that as well.
He then was really fortunate and got onto the abiraterone Phase I/II trial that I was running, which led to the approval of abiraterone and my running the abiraterone Phase III, and after 3 years on abiraterone he progressed. And at that juncture he told us that he had just been at a funeral for his — one of his cousins who had just died of prostate cancer. And in light of what we knew at that time, in 2009/2010, we sequenced his germline in his tumor, and we found the germline BRCA2 mutation. And we treated him in 2010 on olaparib, and he had an almost 3-year response. This is a post-chemotherapy patient with a BRCA2 mutation to olaparib.
His PSA slowly climbed, but he didn’t progress radiologically, and eventually progressed in the pelvis. And at that juncture we actually were able to treat his remaining pelvic disease with radiation. He had a complete response to that pelvic radiation, which lasted for quite a long time. And he eventually died, free of disease, of a completely different problem, and at that juncture he was in his mid to late 70s but died still free of disease.
DR LOVE: Did you give him carboplatin? I mean you mentioned it. Did he get it?
PROF DE BONO: No.
DR LOVE: Okay.
PROF DE BONO: He didn’t get it. He actually — he actually died of a cardiovascular event.
DR LOVE: So I was going to say, he just —
PROF DE BONO: Yeah. And I suspect the longstanding hormonal therapy may have accelerated his atherosclerosis.
DR LOVE: Interesting. Very interesting.
PROF DE BONO: But he was still in a CR when he died.
DR LOVE: Wow. Incidentally, did you keep the olaparib going when you radiated him?
PROF DE BONO: No. But olaparib does radiosensitize, yeah.
DR LOVE: One other thing before you continue. In an earlier slide I noticed this thing, I don’t think you commented on it, a predictor of response to abi, SPOP.
PROF DE BONO: SPOP is a gene that’s often mutated. It’s a CRL E3 ligase. It results in decreased androgen receptor degradation. And basically it results in high AR protein levels and increases AR signaling. And these tumors, we’ve shown, and others have confirmed, are associated with very long responses to drugs like abiraterone and enzalutamide.
DR LOVE: Very interesting.
PROF DE BONO: SPOP.
DR LOVE: Please continue.
PROF DE BONO: It’s one of the commonest mutations in prostate —
DR LOVE: Really?
PROF DE BONO: — cancer.
DR LOVE: Really?
PROF DE BONO: And it’s found in about 1 in 10 advanced prostate cancers.
DR LOVE: Really?
PROF DE BONO: It associates with a better prognosis.
DR LOVE: Is it usually on a typical NGS report?
PROF DE BONO: Yeah. Yeah, yeah.
DR LOVE: And how about liquid biopsy? Do you see it also?
PROF DE BONO: Yeah. You see it in liquid biopsies, yeah.
DR LOVE: Wow. Really interesting. Great. Please continue.
PROF DE BONO: Now I do think it’s important to note here that we have published before, and others have confirmed, both in the lab and in the clinic, that platinum-based chemotherapy and PARP inhibition share similar mechanisms of action. And that action there is cross resistance between PARP inhibition and platinum, although this is incomplete.
What I mean by this is that a patient with platinum-resistant or -refractory prostate — prostate or ovarian cancer, has a much lower likelihood of responding to a PARP inhibitor. And for example in ovarian cancer, in carboplatin- or platinum-refractory ovarian cancer, the PARP inhibitor response rate was less than 5%, whereas in platinum-sensitive ovarian cancer the response rate is 65% to 70%. And in platinum-resistant ovarian cancer the response rate is somewhere in between, 40% to 50%. So this is an important point, that platinum and PARP work through similar mechanisms of action.
Now let me tell you about a second patient. This patient is very interesting. He was a man who was diagnosed in his 40s, very young, he had a very strong family history, and he came to me dying of advanced prostate cancer with advanced liver metastasis and abnormal LFTs. He was getting jaundiced. He had very aggressive disease, was getting a lot of pain. From what you can see here, a large lytic, destructive metastasis in his ilium and his pelvic bone, as well as extensive liver disease. He had had all available therapies for his prostate cancer.
And he was begging me to get therapy, and I knew there was no time here to wait for his genomics; he was progressing very quickly. And because I knew he had a strong family history we proceeded to give him platinum-based chemotherapy. He was not eligible for a clinical trial.
This man essentially had a complete response of his liver metastasis from his metastatic castrate-resistant prostate cancer. And in fact, his tumor regressed so well that actually his bone healed. He had a calcified residuum in his iliac bone, but he had full resolution of his pain. Although actually, he actually, I recall, went on vacation with his kids, I stopped his carboplatin, and his grandkids, to Paris, Disneyland, I recall.
And then interestingly, after he’d been off the carboplatin for some months progressed in a solitary cerebellar metastasis, and as you can see here some recurrence in his liver. And he’d been off carboplatin for some 8 months.
So I gave him carboplatin again, and I was really delighted to note that his cerebellar met regressed on carboplatin alone. His liver met completely regressed, and we then gave him stereotactic radiotherapy to the brain, to the solitary cerebellar metastasis.
He had further long remission from his disease, eventually relapsed again, got platinum again. Now this gentleman died eventually after really about 5 years of disease control from when he came to me dying of advanced prostate cancer. And what was fascinating to me is that this man’s tumor was sequenced in many laboratories; in Ann Arbor, in Seattle, by my friends, Arul Chinnaiyan and Pete Nelson’s laboratories, through Colin Pritchard in Seattle, who were never able to find a mutation in a DNA repair gene.
But actually, when we looked at his HRD score, when looking at his exome sequencing, we found that he had one of the highest-ranked HRD scores of all the CRPCs we had tested by 3 different HRD scores. So this man actually clearly had a tumor that had a homologous recombination defect, but we could not detect this by standard exome sequencing modalities. And it may be that he had a rearrangement of his BRCA genes, for example, through an intronic region that we cannot detect by standard mechanisms. And this is quite important because sometimes you cannot detect these alterations by standard sequencing methods.
DR LOVE: What do we know in terms of efficacy of platinums for this specific strategy, carbo versus cis?
PROF DE BONO: We don’t know anything.
DR LOVE: Any guesses?
PROF DE BONO: For prostate cancer we don’t know. We have data for ovarian cancer, although it’s very small and patchy, but to be honest we don’t know. I would suspect that they’ll both be very similarly active.
DR LOVE: And then —
PROF DE BONO: But obviously carbo is better tolerated.
DR LOVE: Right. Yeah. In other situations, chemo — specific chemo, like lung, whatever, sometimes they prefer cis, that’s why I was curious.
PROF DE BONO: Yeah.
DR LOVE: General oncologists are always asking us why don’t the other tumors do what they do in ovary, which is give platinum and then use PARP maintenance; in breast, prostate. Well I mean I guess — I guess pancreas sort of does that.
So any thoughts about that? Do you think that that’s a strategy that it’s too late to get into it, or is it something worth thinking about?
PROF DE BONO: Well first of all, platinum’s never been approved in prostate cancer and currently it’s not standard of care. I going to talk about platinum next. I think it’s important to mention platinum.
But the other thing is that this molecular subset is much less common in prostate cancer than it is in ovary. I mean BRCAness in ovarian cancer is quite common.
At present what we know is there’s a subset of prostate cancer DNA repair defects, and you have to select for that subset, based on my knowledge of this biology.
Now I’ll come to something quite different later, but at present platinum is not a standard of care.
DR LOVE: One other thing I was just going to ask you about, just something I was going to ask you in general, do you know anything about NRG fusions? And do they occur in prostate cancer?
PROF DE BONO: I know a lot, yeah. They’re very rare in cancers overall. In fact, I’m running a trial of a HER3 targeting antibody for which we’re deliberately pursuing NRG fusions. I’m currently running that trial. It’s a really cool antibody. It prevents HER2 dimerization.
DR LOVE: Is that the same one that was just reported at ASCO?
PROF DE BONO: No.
DR LOVE: A different one.
PROF DE BONO: No. Different one. In prostate cancer, and in colorectal and pancreas, and even in lung, the frequency of these rearrangements of NRG1 is less than 1%, generally. So we are looking for those, and in fact these tumors with NRG1 rearrangements have very high NRG1 protein expression. So you can do a cheaper way to screen the tumor is NRG1 immunohistochemistry.
But more importantly in prostate cancer we see a lot of HER3 expression. So if you look at my paper published last year in Cancer Research, what is particularly interesting in prostate cancer is actually when you treat with hormone therapy you induce the movement of white blood cells into the tumor. And these white blood cells secrete high amounts of NRG1 that drive hormone resistance.
DR LOVE: Wow.
PROF DE BONO: And that paper’s by Gil et all, Veronica Gil et al. She was my postdoc.
DR LOVE: I ask about NRG they haven’t even heard of it. So I’m not surprised that you have. But just trying to understand it.
So these are fusions to a ligand? I mean biologically why is it — is that what it is?
PROF DE BONO: Okay.
DR LOVE: It’s like a fusion of a ligand?
PROF DE BONO: So NRG1 — NRG1 is neuregulin 1, neuregulin 1, okay? Neuregulin 1 binds to HER3 receptor which obviously heterodimerizes with HER2 and EGFR, yeah?
DR LOVE: Got it.
PROF DE BONO: So HER3 signals through the HER2/HER3 or the HER3/EGFR axis. Now what I’m saying to you is that in some cancers a rearrangement happens that causes very high expression of NRG1 in the cancer, okay? But in prostate cancer that’s rare, less than 1%. But despite that, many prostate cancers become addicted to NRG1, but they don’t get it through a rearrangement, they get it by sucking in white blood cells that make it for them.
DR LOVE: Wow. The agent you’re studying, and this other one from ASCO, are they basically monoclonal antibodies to HER3?
PROF DE BONO: Correct. They’re monoclonal antibodies, but what’s cool about the antibody I’m developing is it’s a very potent blocker of heterodimerization. Not all antibodies are the same.
As you may remember, pertuzumab that I studied 20 years ago in prostate cancer, which is obviously approved for breast cancer, is meant to do the same thing but wasn’t very potent. It was meant to block HER2 heterodimerization.
DR LOVE: So let me ask you just one more thing about this because that other thing about it affecting androgen sounds really interesting. This blocking of the dimerization, should that theoretically deal with that?
PROF DE BONO: Correct.
DR LOVE: This androgen activation or something?
PROF DE BONO: Correct. So we are running trials now with the anti-HER3 antibody that includes prostate cancer.
DR LOVE: Wow. And the thing about being less than 1%, that doesn’t stop the lung people at all. They’re looking for —
PROF DE BONO: Yeah, no. It’s an extra. So the NRG1 rearrangement is so rare — I mean it’s less than 1% of all cancers, so it’s very rare. And it includes lung, prostate, pancreas, colorectal, but less than 1%.
But what I’m saying to you, a much bigger deal is that these tumors can actually easily get the NRG1 —
DR LOVE: Right.
PROF DE BONO: — and lots of it from white blood cells. Not just prostate, I’m sure —
DR LOVE: That is so interesting.
PROF DE BONO: — but prostate is the posterchild.
DR LOVE: That is so interesting.
Incidentally, I do these interviews with oncologists in practice. The other day I interviewed
— they had — a doc who had 2 NRG lung cases in his practice. I couldn’t believe it. Most of the investigators don’t have any.
PROF DE BONO: I’ve just sequenced a thousand patients for NRG1, so I have a number because I’m looking for them.
DR LOVE: Awesome.
PROF DE BONO: So yeah, we have several.
DR LOVE: That’s very, very interesting. Please continue.
PROF DE BONO: Now what’s important to remember is as we’re seeing these responses to PARP inhibition and BRCA cancers, including prostate cancer, that data were being published by Sternberg et al, a Phase III trial led by Cora Sternberg and Oliver Sartor, of a platinum actually generated by chemists at my institution called satraplatin. Now this trial sadly failed to lead to registration of the platinum in prostate cancer. In fact sadly platinums were never able to approved for prostate cancer patients.
But actually Cora Sternberg et al in JCO presented these data and published them and reported that platinum and a 25% PSA response rate, just platinum in prostate cancer. Now I would say this is a gentle, weak platinum. But basically these data, to me at that juncture 13 years ago or so, made me realize that a proportion of prostate cancer must have a DNA repair defect leading to platinum sensitivity.
And that the reason this trial failed to show a survival benefit, on the right, is because there was no patient selection. That was my interpretation of the data at that juncture.
So this led me to design a trial that has led to a change in the field called TOPARP. But let me first of all explain to you where I was at at that time. We had shown through several trials, that I led most of them, that multiple PARP inhibitors, including olaparib, niraparib, talazoparib, and rucaparib, had similar antitumor activity with mainly hematological toxicity, most of that was anemia, with antitumor activity only in tumors with a germline BRCA deleterious mutation, including metastatic CRPC, but also sporadic tumors without a germline mutation that have BRCA loss in tumor DNA but not normal DNA. And actually not germline DNA. And we saw activity in these ovarian, breast, prostate, and pancreatic cancers.
We have, as I said, associated platinum-sensitivity with PARP inhibitor sensitivity and published that in a paper by Ang et al in 2013. And based on all this we therefore hypothesized that a proportion, at least 20%, 30%, of advanced prostate cancers would have BRCAness and be sensitive to PARP inhibition.
So I designed this investigator-initiated trial through grants from the Prostate Cancer Foundation, Prostate Cancer UK, and Cancer Research UK, which was an adaptive trial with a test set and a validation set. And the main goal of this trial called TOPARP was to identify prostate cancers that actually would respond to single-agent PARP inhibition. In this case it was olaparib.
So basically in the first part of this trial we treated almost 50 patients, 50 patients actually. One of them had a fracture after 2 weeks of therapy and had to go off trial, so it was 49 evaluable patients. And we showed that all the responding patients, really on the TOPARP they — pretty much all of them, bar one, had an identifiable DNA repair defect.
We then went on to do a validation set to actually now this time molecularly select patients for the TOPARP-B trial. So let me describe these in turn.
This is the first trial that we published in the New England Journal of Medicine from my former PhD fellow Joaquin Mateo, and Joaquin is now a senior investigator in Barcelona. And essentially this is the main figure in paper, and basically what is shows is that the responders to this single-agent PARP inhibitor olaparib, on the left, almost all had an DNA repair defect bar one. And the patients with tumors that didn’t have DNA repair defect mainly did not respond to olaparib of these 49 patients.
Now what shocked us from this trial, in these 49 patients, was that actually, as you can see asterisked, 6 of the 50 patients, that is 12%, had a germline DNA repair defect. That’s why there’s an asterisk. And what’s also interesting is that you see a small number of patients have a mismatch repair defect in this trial, as outlined, for example, here by MLH3 alteration, indicating that actually rarely in prostate cancer you do get, and here’s another patient with MLH3 alteration, you do get rarely mismatch repair defects in this disease. As I said earlier, it can sensitize to immunotherapy.
And what was important for us was that actually these responses were quite impressive. 90-percent-plus falls in PSA in men that have no therapy options left. Durable partial responses, falls in circulating tumor cell counts, and obviously of course RECIST responses. And the main toxicity from this drug was anemia. And then in a small number of patients the patients had to have a dose reduction, although a very small number, also those that had to discontinue because of myelosuppression. These are obviously very late-stage disease patients.
Now we then went on to show from this work, by sequencing all 700 patients across our Stand Up To Cancer and International Prostate Cancer Dream Team, germline mutations in mCRPC and show that 1 in 7 men, essentially, with metastatic CRPC had a hereditable deleterious mutation in germline DNA, exactly like the TOPARP-A data, 12% or so had a germline mutation. If you had a family history the risk was higher, probably about 20%. And if you lived in certain cities, like London and New York, you had a higher likelihood based on the demographics and the pharmacogenetics of our populations.
And this changed the NCCN Guidelines internationally, and all men with a metastatic CRPC are now getting and being recommended to get germline DNA testing for DNA repair defects. The commonest defects being BRCA2 and ATM mutations, as you can see here. PALB2 is also an important gene. BRCA1 alterations are much less common.
So family cascade testing is also now indicated when these men have detectable alterations. Obviously the goal here is to prevent the family members of these men getting cancer through preventative mechanisms, although we need to figure out better ways to prevent prostate cancer and breast and ovarian, pancreatic, and other cancers in these patients, including not only surgery, but hopefully preventative strategies.
Now that trial that did not require molecular preselection led to the validation set, which we published in Lancet Oncology just over a year ago. And this was where we treated almost a hundred patients with metastatic CRPC, but the patients were all found to have at least 1 allele, 1 gene copy altered with a DNA repair defect. Again, this was an investigator-initiated trial run across the UK.
So this was the validation set. The tests had shown us that the patients that responded mainly had DNA repair defects. This validation set went on to confirm what this activity looked like in these preselected patients.
And these were the results here. The BRCA patients did very well, had really impressive and durable responses.
And importantly most of the patients that had 1 BRCA alteration on deep genomic testing, almost always, 80% to 90%, had biallelic BRCA loss. So if you had 1 allele hit on the BRCA you almost always had 2 allele hits on BRCA.
I should say this, however. We saw other alterations sensitized to PARP inhibition, not surprisingly, including PALB2 alterations. They had to be biallelic, and I’ll come back to that later, but again this resulted in profound responses to PALB2 biallelic loss, both biochemically and radiologically, that were often durable. And many of these PALB2 alterations were actually germline alterations.
I should also, however, point out that the dark green alterations are the ATM alterations. And you can see there are patients with dark green ATM loss responding to PARP inhibition single agent. So ATM loss can sensitize, particularly when it’s homozygous deletion ATM loss, and I’ll come back to that later on. And these are usually durable responses.
So these are the data here, again same color scheme. You can see the patients that did best for rPFS and duration of response are the BRCA patients, but the ATM patients can also have more than 6-month disease-free control, which in this late-stage population is very important. The PALB2 patients are here in light green. Again, several patients weren’t on the trial for more than 6 months. And interestingly we see some evidence of disease control, even in the CDK12 alterations, although this is much less common. And some other alterations, for example FANCA alteration, biallelic loss can sensitize to this single-agent PARP inhibition strategy; and look in the paper for more detail.
So here’s a PALB2 biallelic loss patient who had a good radiological response in his neck node, as you can see I wrote here, lasting really for over a year. And you’ll see that actually the devil is in the detail. BRCA2 loss, or BRCA loss, does sensitize to PARP inhibition, but it’s very important to know that the BRCA2 HOMDELs, where both BRCA2 genes are deleted, result in approximately threefold longer disease control than when you have BRCA2 biallelic loss with 1 allele being lost by mutation. And the reason for that is the cancer can relatively easily restore the BRCA2 gene through a second mutation that reverses the BRCA2 back in frame and restores BRCA2 function, which causes resistance to PARP inhibition. Whereas if you have BRCA2 homozygous deletion it is much harder for the cancer to find a resistance mechanism.
Now you’ll see here the swimmer plots for these tumors, and importantly you’ll also see the biallelic loss, identified by the blue bars on the left of the graphs, and you’ll also see that the germlines are the darker shade for both patients. And most of the germlines, as you’d expect, have biallelic loss. And what’s important to note is that the patients that do best with PALB2 alterations have biallelic loss, and most have germline alterations, and the patients that do best on — with BRCA2 loss have homozygous deletions, as I was saying earlier. Please note that’s really quite important because the BRCA2 HOMDELs are very easily missed if the tumor DNA sequence is of low allele frequency, as is common in plasma DNA.
Now similarly for ATM loss the patients that do best have biallelic loss, as demonstrated in these survival curves here, the rPFS. And the patients that do best with ATM loss probably have ATM homozygous deletion, as depicted in these figures here.
And as expected, ATM loss does not cause HRD, and therefore, as you see here, this is a figure that looks at the γH2AX and the RAD51 foci assay. RAD51 foci indicates the absence or presence of HRD (homologous recombination). If you have no RAD51 foci, as is happening here with the BRCA biallelic-loss tumors, no RAD51 foci you have HRD. If you have RAD51 foci, as you’re seeing with ATM, in dark green, you don’t have HRD. But ATM loss loses the sensitization to detecting DNA damage, so it’s indirectly causing HRD. So CDK12 loss and ATM loss do not cause HRD directly.
But you’ll see that the PALB2 loss, biallelic, loss, does cause HRD. And the others are mainly non-HRD inducing. So important points. This is a true immunohistochemical functional assay for HRD loss by doing immunohistochemistry for RAD51 and γH2AX.
Now my investigator-initiated TOPARP trial led to a trial that I designed called PROfound, and this is a trial studying single-agent olaparib in men with pre- and post-chemotherapy mCRPC, preselected to have DNA repair defects. The trial was primarily powered on Cohort A, with Cohort B really being a pilot study, Cohort A treating patients with BRCA and ATM alterations. I should be very clear, this is monoallelic. We did not have to require biallelic loss here. And Cohort B requiring a number of different DNA mutations, including most importantly genes like PALB2, RAD51, et cetera, although the commonest gene in that cohort is actually CDK12 alterations.
The patients could have had no prior chemotherapy or some chemotherapy, 1 prior chemotherapy, and the randomization was to olaparib single agent versus physician’s choice with crossover at progression on the control arm, very importantly. One of the very first, if not the first, positive Phase III trial in prostate cancer that allowed the control arm to the treatment arm.
The primary endpoint was rPFS, and key secondary endpoints in a hierarchal fashion included, as you can see here, confirmed radiological response rate, time to pain progression, and overall survival.
And we published in New England Journal of Medicine there was a significant, and this met the prespecified criteria, improvement in rPFS, with a p-value of less than 0.0001 and a hazard ratio of 0.34, that is a 66% reduction in the risk of progression versus physician’s choice. And when we did the final survival analysis we confirmed that, as you can see here, overall survival improved, which was surprising, despite two thirds of the patients crossing over. And when you did the prespecified adjustments for crossover the hazard ratio revealed 0.42, and a p-value here, as you can see, that met the prespecified cutoff. And this led to this drug’s approval for advanced prostate cancer with not only the BRCA and ATM for Arm A, but also the other DNA repair defects.
Now we went on to present data regarding response rates of olaparib in this trial. And you can see that olaparib is causing response not only in the BRCA, in the PROfound population, but also you see responses in the ATM population, I’m sure primarily in the biallelic loss tumors, some uncommon response in the CDK12 loss tumors, but actually most of the benefit is from the BRCA and ATM loss cancers, the most commonly aberrant tumors.
Now I should say we are seeing similar data with other PARP inhibitors. And this is the rucaparib data from my friend and colleague Wassim Abida, and you can see that rucaparib, a very similar PARP inhibitor to olaparib — this was only, though, in the post-chemotherapy setting, only in BRCA-altered cancers. And again you can see a really impressive response rate by PSA and by RECIST criteria.
And here is data that I led for talazoparib, arguably the most potent PARP trapper, and again you can see responses not only in the BRCA patients with a PARP inhibitor, but also, as you can see in orange, in the ATM-loss tumors and also in some of the other cancers, including PALB2-loss tumors, in red.
ATM has been a major issue for me because some have argued that these tumors don’t work in ATM loss tumors. But let me remind you that actually in the New England paper that we published, we had 2 papers with — we had 2 patients, sorry, with germline mutations of ATM. These patients had biallelic ATM loss, ATM IHC loss, and what you can see is duration of treatment for over a year in 3 patients with ATM alterations and major PSA falls, 85% fall in PSA, 51%, 59% that was durable, with CTC count falls.
So there’s really no doubt that actually ATM does sensitize, although beware in ATM mutations, particularly from blood, is not necessarily ATM loss. We published this in this paper by Neeb et al in European Urology. In fact, most of the ATM alterations, SNVs (single nucleotide variants), do not result in ATM loss. So beware. And I fact I would say that if you’re doing plasma it’s even more complicated because you can have clonal hematopoiesis mutations in the hemopoietic progenitors that actually are not in the tumor that can mislead you further.
In conclusion, synthetic lethal strategies have changed in prostate cancer care, but actually, remember, this is not just PARP inhibition. If you cannot get the patient a PARP inhibitor consider carboplatin, for example AUC 5 or 6 three weekly. Based on our work germline DNA repair testing is now standard of care for all men with metastatic CRPC, advanced disease. And remember, this will result in family cascade testing. And always take a family history from a man with advanced prostate cancer, and if that man has a family history his chances of having a DNA repair defect as germline is probably at least 1 in 4 to 1 in 5.
We have an increasingly complex predictive biomarker landscape, but remember, biallelic loss of the DNA repair gene is necessary to sensitize the PARP inhibition and platinum. But importantly, if you have loss of 1 allele of BRCA2, in 80%, 90% you have biallelic loss. That is not the case with ATM, but ATM loss, biallelic loss, does sensitize to PARP inhibition. And beware of the risks of plasma ctDNA NGS. The median tumor fraction of plasma DNA, that is most of the DNA in plasma, is not tumor DNA, and this almost always misses the HOMDELs unless you have special assays like low-pass whole genome sequencing. And beware that the clonal hematopoietic mutations can confuse the field, particularly for ATM.
Now, finally, it has been argued that PARP inhibition may work in tumors that do not have DNA repair defects. And the reason for this has been argued that PARP inhibition blocks AR signaling or maybe the AR blockade downregulates DNA repair genes like BRCA and sensitizes to PARP inhibition. But let’s be clear, we have not seen PSA falls in men with prostate cancer that do not have a DNA repair defect. Moreover, if abiraterone did cause HRD you would have major sensitization to PARP inhibition, with a huge increase in response rate, as seen in the BRCA patients, 70%, 80% response rate. We have not seen this in either PROpel or MAGNITUDE.
So when you look at the PROpel and MAGNITUDE data what you have to consider is A) is this primarily being driven by the DNA repair defective population, or B) does PARP inhibition have some other mechanism of action.
DR LOVE: That was really great. If I could just follow up with a couple of questions. First, getting back to your patient who got the carbo that had the very high LOH. First of all, I only kind of heard like about — like over or under 10, like how high does LOH go, and what level is it that you start thinking that’s like really high?
PROF DE BONO: I don’t know.
DR LOVE: Okay.
PROF DE BONO: Basically, the truth is that actually cancer is so complex that cutoffs are just very arbitrary and artificial. And particularly when — as is likely you are dealing with cancers that are in 1 patient heterogeneous, with some tumor having HRD and other cells in the same patient not having HRD. So setting a bar is very artificial and highly risky, in my opinion.
The other problem with prostate cancer is that unlike ovarian cancer, for reasons that I don’t fully understand, the HRD scores don’t work that well, and I don’t know why that is.
DR LOVE: In prostate you’re saying.
PROF DE BONO: In prostate.
DR LOVE: You’re saying in prostate.
PROF DE BONO: So in breast cancer and ovarian cancer, but especially ovarian cancer, the HRD scores are superb. In both breast and ovary the HRD score really identifies the patients with DNA repair defects. In prostate that does not seem to be the case.
DR LOVE: So how do you explain what happened with that man?
PROF DE BONO: Well, I think that man probably had a DNA repair defect that we have not been able to detect. Maybe, for example, an indel that actually can be very hard bioinformatically to identify. And what I can tell you is that because I was sure there was something I sent this man’s DNA for sequencing to Arul Chinnaiyan’s lab, outstanding, award leading lab —
DR LOVE: Yeah.
PROF DE BONO: — to Colin Pritchard’s lab, CLIA certified in Seattle, superb, leading group. And obviously we did it in our lab by multiple different mechanisms, deep whole genome sequencing. We have never found anything. So basically I don’t know why he responded.
DR LOVE: I don’t know whether you saw the paper that was just presented and then published at ASCO on the ATHENA trial of rucaparib in ovary.
PROF DE BONO: Yup.
DR LOVE: And there it was again, HRD proficient. I know it’s a different cancer, but again, do you think that maybe through some other mechanism or there’s some mutations that we’re not aware of that —
PROF DE BONO: For sure.
DR LOVE: — that in fact — from a clinical point of view right now do you think clinically with what a doc has in practice they can identify patients who are not — at least let’s say in the primary maintenance setting, can you identify people who are not going to benefit by PARP inhibitors?
PROF DE BONO: So let me remind you about that patient that was particularly important. 1) He had a very strong family history. 2) He was young when he was diagnosed with metastatic disease. He was 44. And 3) he had metastatic liver disease that often makes me think. If I see certainly unusual disease in the liver or say the lungs I always think this is not a typical prostate cancer. Something else is changing. So I think there was something quite different about that man that basically made me highly suspicious, particularly the family history, that he had an HRD defect, which is why I gave him platinum.
So I gave him platinum before I had the HRD score because I suspected that he would respond, and he obviously did.
DR LOVE: Yeah. Well that — certainly you can’t argue with that, although it’s, again, kind of wondering exactly what’s — why.
So again, PROpel, I think it was 0.54 or something hazard rate, HR proficient. So how do you explain it?
PROF DE BONO: But remember the satraplatin trial for PFS had an almost identical hazard ratio, okay? So actually it’s maybe for PFS. So it may be that the DNA repair defects, and I’m not saying it is, but maybe the DNA repair defect — DNA repair-defective tumors drive the hazard ratio. And — but I think, but be honest, there’s still questions to ask here. One thing that’s key about PROpel is that it is very likely that the — what they have called HRD proficient includes false negatives. And as I said very clearly in my talk, the super responders have BRCA2 HOMDEL, and their assay definitely misses them.
DR LOVE: So I mean there’s the theoretical and translational aspect of this, which you’re deeply into, but there’s also the practical aspect.
PROF DE BONO: — good trials will always lead to more questions. I think the key thing is that we have advanced the field. We are stratifying the disease. I think there is no doubt there’s a lot more to learn. There is no doubt that we are missing patients that can benefit. But the question remains do you — do you — as I called it, are we splitters or are we lumpers. Do we treat them all the same?
DR LOVE: Absolutely.
PROF DE BONO: Or do we treat them in different ways? I think at present I have to be a splitter.
DR LOVE: And there’s a huge lumping and splitting thing going on in oncology in general about PD-1.
PROF DE BONO: Yes, sir.
DR LOVE: Like you see — you see trials where everybody benefits, but then you pull out the PD-1 negative and they don’t. Nobody knows what to do. So that’s like a huge issue.
PROF DE BONO: It’s a lot worse for PD-1. I mean one of my friends just got PD-1 therapy, and she’s had horrendous toxicity. Young woman with melanoma, and honestly, she almost died from toxicity. She was in the hospital for weeks.
And I think you have to remember olaparib and PARP inhibitors can cause acute myeloid leukemia. If we start giving these drugs early in the disease setting, and the patients are on them for years on end, these guys will end up with leukemia. I mean I have no doubt that will be the case. So my concern here is we have to as physicians remember our mandate is do no harm.
DR LOVE: One final question. Putting aside reimbursement and regulatory issues, right now what’s your inclination in terms of adapting the PROpel and MAGNITUDE data? And in particular if you could, would you use PARPs, I guess specifically olaparib, in somebody who’s HR proficient?
PROF DE BONO: I think based on the data we have I think I could not recommend olaparib for all comers. I think more data will come from PROpel and other studies, and we will understand this better. But at present I cannot really recommend olaparib for everybody, at present.
And the reason for that is that when you give olaparib with abiraterone you really have no idea what’s working. Is it the abiraterone or the olaparib? So if your PSA is going away — and with abiraterone you can be on it for years, right, 4 or 5 years, and my concern here is do no harm, if you’re going to give a patient olaparib and abiraterone for many years, then you have to be jolly well sure that that patient’s going to get benefit. And remember in my TOPARP trial we did not see any PSA falls in the men that did not have DNA repair defects.
DR LOVE: So again, just getting into the practical issue then, so again, assuming you have a patient who has no — HR proficient, and I hear you saying that you’re hesitant. I guess my question would be would you have a different viewpoint if in that subset — I think the hazard rate was 0.76, if it were like 0.3 would you feel differently?
PROF DE BONO: I think if we showed a significant overall survival advantage in tumors that were definitely HRD proficient. But remember, I am certain that what they’re calling HRD proficient in PROpel has false negatives. And indeed, my colleague Neeraj Agarwal raised this issue after the presentation for PROpel at GU ASCO and said well what about the risk of false negatives in this population. And I fully agree with him. There’s a significant risk that this trial missed the BRCA HOMDELs, particularly because a significant number of patients had plasma DNA sequencing, but also because in prostate cancer all biopsies have poorly-preserved DNA.
DR LOVE: I’m just trying to figure out from a practical point of view how much of your concern is based on translational understanding and how much is it based on the actual clinical data.
Like 0.76 —
PROF DE BONO: Remember, PROpel has not shown — it’s not shown an OS benefit. PROpel has not shown a survival benefit.
DR LOVE: I mean osimertinib hasn’t shown a survival benefit either, but the hazard rate’s 0.2, and the hazard rate for brain mets is 0.1. But that’s my point. Like maybe —
PROF DE BONO: Would you give osimertinib to all lung cancer patients? No.
DR LOVE: No, but I’m just saying that based on the PFS data they wanted to go ahead with it. It’s a different disease, different situation —
PROF DE BONO: Yeah. But for osimertinib it makes perfect sense. The drug works in that subset. This is different.
DR LOVE: Yeah. I understand the translational basis is different, but again, I’m just thinking about docs in practice trying to make decisions.
PROF DE BONO: Right.
DR LOVE: All I’m saying is when you have trials in general, in solid tumors in general, if you have trials that are kind of early so all you have is PFS benefit, and the hazard rate is 0.76, people are kind of plus/minus, am I going to do that without survival benefit. So they feel differently when they see it —
PROF DE BONO: Yeah. The question for us today is this hazard ratio largely driven by the HRD tumors, that’s the question, including the false negatives in their HRD proficient cohort.
DR LOVE: Well again, that’s a research issue. But what I’m talking about is the practical issue of when do you use a therapy with the data that you have.
PROF DE BONO: Well let’s see what the regulators say. We’ll have to wait on the FDA and the EMA. As you know, the EMA were very stringent. They only approved olaparib for BRCA, which I did not agree with. I begged the EMA to approve olaparib for ATM as well. But we’ll have to see what the regulators say. I can’t comment.
But today what I know for certain is that olaparib works primarily in the HRD tumors and most impressively in the BRCA2 HOMDELs.
Available Data with, Ongoing Investigation of and Potential Future Role of PARP Inhibitor-Based Combination Strategies — Fred Saad, MD
DR SAAD: I’m Fred Saad, Professor and Chairman of Urology at the University of Montreal Hospital Center here in Montreal, Canada.
So the prostate cancer landscape is really quite vast, but really when we think about it and when we look at our patients almost all the suffering and all the death from prostate cancer occurs in the metastatic castration-resistant prostate cancer state. Until we reach this state patients really don’t suffer from prostate cancer, and especially don’t die from prostate cancer but rather with prostate cancer. And ultimately we’d love to see this state of the disease disappear by introducing therapies earlier that allow patients to remain nonmetastatic castration resistant until they die of something else. But until then we really need to work on improving treatment options for patients with mCRPC.
So metastatic prostate cancer, we realize, as patients progress become and more heterogeneous, and multiple pathways for resistance occur, of which about a quarter of these pathways are driven by mutations in DNA repair pathways. And there have been studies looking at patients with germline mutations with metastatic CRPC and about 1 in 10 harbor germline mutations that are pathogenic. But if we use only germline testing alone about 50% of patients with mutations would be missed in terms of somatic mutations with BRCA1/2 or ATM.
Why these mutations are important is there’s a lot of work showing that HRR mutations, including BRCA2, which is the most common, have poor outcomes on standard of care therapies and in general have worse survivals than patients that are metastatic castration resistant without BRCA2 mutations. And here the example is survival in noncarriers being about twice as long as carriers of BRCA2 mutations.
So a lot of work since 2015, and PARP inhibitors in prostate cancer have exploded, basically, and now we have multiple treatments being investigated, and we’ve already got FDA approvals for olaparib and rucaparib based on Phase III and Phase II studies, and very encouraging for the future.
DR LOVE: I was just reflecting back on what you were just saying about the prognosis of patients with, for example, BRCA2, and I think that actually in ovary, which I know is a different disease, I think they actually have a better prognosis. Maybe that’s because they get platinum therapy or something. I don’t know. But what’s the thinking about why their prognosis is worse in prostate?
DR SAAD: Yeah. So obviously I’m not an expert in ovarian cancer, but I think one of the reasons they might have better outcomes is they have an actionable target for drugs that are actually effective, whereas in prostate cancer we have many effective drugs for patients with prostate cancer, especially metastatic CRPC, but we’re just starting to tap into an actionable mutation that can make a difference. And so these patients until recently were not treated any differently than the average patient, so their survivals were worse because we really didn’t have anything to offer those patients. So I think that’s why there might be a difference.
In ovarian they probably don’t have as many good therapeutic options as we have in prostate cancer, so they’ve addressed BRCA mutations in ovarian adequately, but they still need to work hard on patients who don’t harbor these mutations, whereas it’s the opposite in prostate cancer. Most patients don’t have these mutations.
DR LOVE: Yeah, no. I mean definitely there’s no hormonal therapy or certainly not a major thing in ovary, but I guess the thing I’m wondering about is, is there any connection between BRCAness, that whole thing, and the androgen pathway. Is there a difference in how the androgen receptor or the androgen pathway works?
DR SAAD: Yeah. There’s work going on to look at whether the pathway actually is helpful when we’re talking about combination therapy that I’ll talk about in a few slides. But clearly there’s a lot of work left to be done, and we need to address this small group of patients that really need to do better than what we’re able to offer the average patient with prostate cancer.
DR LOVE: Interesting. Please continue.
DR SAAD: So I think people are quite aware of the PROfound data that was presented and published in The New England Journal of Medicine, the first randomized Phase III study with a PARP inhibitor in metastatic CRPC that took patients that harbored mutations and had progressed on at least 1 line of novel hormonal agent, like abiraterone or enzalutamide, and were randomized to olaparib versus another novel hormonal agent. Patients were allowed to cross over when they progressed if they were on the control arm. And the results show that we were very effective in delaying radiographic progression and improving overall response rates in patients that got olaparib.
But the question is, and this is some of the rationale to going to earlier settings, is earlier better with a drug like olaparib in these patients. And clearly the first evidence is the overall survival improvements that were seen in this study, and patients had a 31% reduction in death if they got olaparib versus another novel hormonal therapy. And this is in the face of being allowed to cross over to the olaparib arm. So really allowing patients to cross over did not allow them to catch up in terms of overall survival, and those improvements would have been actually a 58% reduction in the risk of death if patients hadn’t been allowed to cross over. So at least some evidence that maybe earlier treatment would be more effective.
And so really where we’re at now is looking at earlier introduction of PARP inhibitors, looking at combination trials, and even at addressing do we need to limit PARP inhibitors only to patients who harbor HRR mutations. And this is what I’ll focus on in this part.
So just taking a step backwards, we have had multiple Phase III trials that have been successful in improving overall survival in mCRPC. Very significant improvements, but still relatively modest, in the range of 4 to 5 months, in about — all of these trials. But one striking feature of all of these trials is that the control arm of all of these trials had a non-life-prolonging control arm, either a placebo or a switch in hormonal therapy that we’ve never shown was able to improve overall survival. So we still have to struggle with improving outcome over a well-described control arm that is life prolonging.
So one of the most commonly used therapies for mCRPC is abiraterone, and we just want to remind what kind of results we obtained with abiraterone against placebo. We were able to get 16.4 months rPFS compared to 8.2 months with placebo, so about an 8.2-month improvement in rPFS in patients getting abiraterone compared to placebo. And we were able to improve overall survival by about 4 months against a placebo.
But in the real world we’re not doing that well. And in the real world consistently mCRPC patients are living less than 2 years because many patients really don’t go beyond first-line treatment for mCRPC. So we really need to address improving first-line therapy for mCRPC, which is the most effective, but also for many patients, unfortunately, the only line of therapy they’re going to get.
DR LOVE: What do we know about why people develop progressive disease on ADT.
DR SAAD: Well, I think we’ve recognized that ADT has led to spectacular results ever since Huggins showed this over 60 years ago. However, Huggins was the first to recognize that it’s a noncurative treatment, that patients will eventually develop resistance and die of prostate cancer. So he was the first to actually add this to his Nobel Prize essay.
And so why they develop resistance — the question is why do patients develop resistance in all diseases, right. I mean cancer is intelligent. We’ll find ways to bypass the treatment that we’re giving them, unless we’re taking patients that are early and nonmetastatic, where we have an opportunity to cure. But these cells harbor clones that are probably already hormone insensitive to start off with, and then there are others that will become hormone insensitive as they’re exposed to the stress of ADT and find ways of surviving even in the face of androgen deprivation therapy. So clearly patients will develop resistance to basically anything we give when they’re in the advanced state of prostate cancer.
DR LOVE: So basically your vision is that there’s a hormone insensitive clone that grows out?
DR SAAD: That grows out or is there likely from the start —
DR LOVE: Right.
DR SAAD: — in advanced disease.
DR LOVE: Right.
DR SAAD: And what happens is we’re controlling the clones that are hormone sensitive, but we’re allowing the clones that are hormone insensitive to gradually take over. And this is a lot of the rationale of why we’re introducing therapies even before castration resistance, right. In hormone sensitive disease we’re starting to use combination therapies, doublets and now even triplets, to try to address this clone of hormone-insensitive cells that are probably there from the beginning. And by addressing that very aggressive clone up front we might do better. And the results have actually shown we do better when we treat aggressively in the first line rather than waiting for resistance and introducing therapies in a sequential fashion.
DR LOVE: I mean I guess, really, that’s what was seen in breast cancer, if you think about it. It’s not that much different in breast cancer in terms of the fact that, like you said, it’s generally not curative in the metastatic setting, but if you move it into the adjuvant setting then you start to see an impact on survival. So I know you kind of had that feeling. You’ve verbalized that to me before about prostate cancer.
DR SAAD: Absolutely. And we’re way behind in prostate cancer or thinking of combinations earlier, but we’re catching up. But there are diseases that treat with quintuple therapy up front, and they actually cure incurable disease. And this was the title of one of my editorials is we’re starting to cure the incurable prostate cancers, and that’s what’s really encouraging by bringing these therapies earlier. But we have to go in a stepwise fashion and prove that it makes sense and that it actually works.
DR LOVE: So please continue.
DR SAAD: Thanks, Neil, because that introduces this part of what the rationale is for combining PARP inhibitors with novel hormonal agents. And there is preclinical evidence, very well-done laboratory work, looking at PARP inhibition reported to increase NHA activity through its androgen-dependent transcription. And that actually the novel hormonal therapy may actually induce a HRR deficiency or an environment that is BRCA similar and increase susceptibility to PARP inhibition, and that this can combined effect would lead to better outcomes rather than using each in sequence and might actually be useful even in patients that don’t harbor HRR mutations.
So this theory from the lab went to a Phase II study that I helped to co-lead looking at olaparib plus abiraterone in patients who had already failed docetaxel in mCRPC. And we looked at radiographic progression-free survival in patients getting abiraterone alone, which was the standard of care, versus abiraterone plus olaparib, and we were very happy to find that we significantly improved rPFS in patients that got the combination over patients that got abiraterone. And when we looked at their HRR mutational status we realized that patients were getting benefit regardless of whether they harbored mutations or didn’t.
So we went from the lab to Phase II and then to Phase III, which is the PROpel study, which we looked at patients, all-comer patients, in first-line mCRPC randomized to getting olaparib at full dose and abiraterone at full dose compared to placebo plus abiraterone, which abiraterone would be considered a standard of care for mCRPC in the first-line setting. We allowed patients to come in who had had docetaxel in the mHSPC setting, and we allowed patients to come in with visceral metastases, and we stratified according to site of metastases and whether or not they received taxanes.
And the primary endpoint, which was discussed with the FDA, was radiographic progression-free survival or death by investigator assessment. And obviously key secondary endpoints that were very important, obviously, including overall survival. However, the study is 800 patients. So clearly with 800 patients against an active control arm, life-prolonging control arm, overall survival cannot be our primary endpoint because all the other studies that showed overall survival almost all had over a thousand patients against a placebo or a non-life-prolonging agent.
DR LOVE: Before you continue, I’m just kind of curious. The other endocrine option that’s often considered in this situation is antiandrogen/enzalutamide. Why was the decision made to just go with abiraterone? And is there any reason to think that abiraterone plus a PARP inhibitor would be different than an antiandrogen and a PARP inhibitor?
DR SAAD: Yeah. That’s a great question. So the decision to go with abiraterone in this study is that we had done a lot of preclinical and clinical work looking at the interaction between olaparib and abiraterone, and with abiraterone we were able to give full dose olaparib.
So even though abiraterone and enzalutamide alone, I would say, are identical in terms of efficacy, in combination there are interactions that lead to more adverse events with one over the other. At least for olaparib we were able to give the full dose. It was tolerated, and we didn’t encounter adverse events that precluded us from using the full dose of both in the PROpel study, which is something we can’t say when we’re using enzalutamide, which is an excellent drug, but in combination with PARP inhibitors oftentimes they have to be reduced in dose.
So looking at the baseline demographics, and I’m only showing 1 that’s the most important and often comes up, is what was the mutational status of the patients who come into PROpel, because this was not a stratification factor, but just by the size of the study we expected it to be well balanced. And so very interestingly HRR mutations were found in 28%, 29% of patients coming into the study. So this is already informative. This is the first study to really take an all-comer population and find that this is the percent of patients that harbor mutations, and that means almost 70% harbored no mutations that were detectable. And by using tissue and ctDNA to determine status we were left with only about 2% where it was an unknown status in terms of HRR mutations, so allowing us to do a lot of work now and in the future looking at the importance of these mutations.
So the primary endpoint of rPFS was clearly met. Amazingly, we were able to get an 8.2-month improvement in rPFS with the combination of olaparib plus abiraterone over abiraterone alone, which is identical to what was seen almost 10 years ago with abiraterone against a placebo, 8.2 months improvement in rPFS, a 34% reduction in the risk of progression or death.
And if we look at blinded independent central review it actually goes up to 11 months or a 39% reduction in progression. So consistently good results in terms of rPFS, and all the subgroups benefitted, clearly, to getting the combination over abiraterone alone. And this includes patients that were young, patients who had visceral disease, patients who had had docetaxel in the hormone-sensitive state, and importantly whether or not they harbored HRR mutations benefitted from the combination over abiraterone alone.
The data is still too immature to talk about overall survival, given the fact that the control arm is an active life-prolonging control arm, but things are going in the right direction, and we should be reporting updates of survival in the next few months.
Time to subsequent therapy was clearly improved with the combination over abiraterone alone, and time to second progression was clearly improved. And so this is all helpful because it points in the direction of consistent improvement with combination of olaparib plus abiraterone.
Reassuringly, the safety profile was very good in these patients in the first-line setting, with only 15% Grade 3/4 anemia, which was very easily managed. All the other adverse events were under 5% in terms of Grade 3/4 adverse events, so very reassuring, and allowed patients to be able to stay on therapy for a very long period of time.
The other study looking at first-line mCRPC in combination therapy is the MAGNITUDE study that actually did things a little bit differently. They took still first-line patients, but they prescreened for biomarker status, and these are the biomarkers they looked at, obviously including BRCA1/2 and ATM and others. And if patients were biomarker positive they were randomized to niraparib at reduced dose of 200 mg because of the interaction with abiraterone obliged us to reduce the dose to 200 mg rather than the full dose. And if they were biomarker negative, then they were randomized for the biomarker-negative group.
The biomarker-negative group did not show any improvement, and so this arm was stopped early, and I won’t discuss that arm because focusing on the biomarker-positive arm.
DR LOVE: Before you go on could I just clarify one other thing?
DR SAAD: Sure.
DR LOVE: Actually, I was trying to figure this out over ASCO because with niraparib, again in ovarian cancer, which has a long history with niraparib, they eventually settled into an algorithm for dose based on platelet count and weight, starting at 300 mg and then they would go down to 200. So with you starting at 200 because of the interaction, if you have a smaller man, I mean for ovary I think it was 77 kg, which is not that small, but I mean a 150-pound man would be under that, would you further reduce it to like 100?
DR SAAD: So the study insisted on starting at the 200 mg dose and not further reduction.
DR LOVE: Really?
DR SAAD: But obviously if patients experienced an adverse event then they had to dose reduce. But the issue with dose reductions is that at some point it becomes less effective, and I think at least in ovarian cancer they showed that at a reduced dose there was slightly less effectiveness. And this is always the worry about reducing the dose in PARP inhibitors. I think in BRCA1/2-mutated patients, even at reduced dose we get efficacy, as was seen here in this study.
But in patients that don’t harbor mutations, or don’t harbor BRCA mutations, you might need a higher dose to get the effect that you’d like to get. And in this study the BRCA1/2-mutated patients, even at the reduced dose, did get a significant improvement in rPFS, with a hazard ratio of 0.53. But looking at the overall group of biomarker-positive patients the hazard ratio was 0.73. And so I think these are all very positive results, but really — it’s really the BRCA1/2 patients that are driving the efficacy in patients getting the reduced dose of niraparib, at least in this MAGNITUDE study. But clearly they are getting a benefit of the combination therapy in the first-line mCRPC setting.
DR LOVE: They did if you look at this. This actually answers my question because it looks like they didn’t get very bad thrombocytopenia there, so I guess the 200 mg works, at least in men, with abiraterone.
DR SAAD: Yeah. Yeah. It works. It works with men at 200 mg. There was anemia at a 30% rate of Grade 3/4, so these are issues that we do have to deal with, and there is a thrombocytopenia of 6.6%, which was about threefold over abiraterone alone. We have to remember, these are patients that harbored mutations, so probably slightly sicker patients and progressing slightly faster. But these are the things that we have to keep in mind, where niraparib does look like to have maybe a safety profile a little bit different than a drug like olaparib. And this is at the reduced dose, so patients, if they experience Grade 3 adverse events, they usually have to even further dose reduce from 200 mg.
DR LOVE: Let me ask you, people hesitate to do indirect comparisons, but sometimes that’s all you have, and some people, I think including you, have looked at MAGNITUDE and PROpel and said well it looks like there’s greater benefit in PROpel with olaparib. But do you think that’s the case or do you think the way they defined it and the mutations they were looking at were a little bit different, and maybe the drugs have really the same effect if you look at the same patient? Or do you think that one is more effective than the other?
DR SAAD: Yeah. So those are very fair questions, and it’s always a bit of a concern doing indirect comparisons between studies. But why I think there might be more efficacy is I think it differs for some patients getting the full dose of a PARP inhibitor. And so we did see improvements in all comer, so the overall study, where 70% we did not find any mutation, if we look at the overall group there was a significant improvement in radiographic progression.
And even if we look specifically at the subgroup of patients without mutations we see a benefit. And within the mutated patients we did not select patients. So only about 30% of patients harbor BRCA mutations, and we had a hazard ratio of 0.5 in the whole BRCA — in the whole mutated group, of which only 30% were BRCA. So there are a lot of things that might indicate that there’s more efficacy, but I’m not sure it’s really the drug itself. It’s the amount of drug we’re able to give and the tolerability that patients were able to take —
DR LOVE: Right.
DR SAAD: — with the full dose. And I think at least in this population full dose looks like it might make a difference when we’re combining agents in the first-line setting.
DR LOVE: That’s interesting. It’s funny because you just presented this a few months ago, but now when you’re presenting it I’m kind of getting a different feel for it. And I’m actually wondering whether either or both of these are going to get approved in this situation, particularly niraparib. I mean I think it seems like olaparib; it would be hard not have that available with these data. But now that you’re going through this, I wonder if there’s really a compelling reason to have niraparib available. What do you think?
DR SAAD: Well I really hope it does get approved, in fact, but I think it’ll be limited to BRCA1/2 patients —
DR LOVE: Right.
DR SAAD: — because the data is very supportive of those patients, and those are the patients that many people are focusing on. And so if we’re — if we know the status, if we’re able to determine the status, and they’re BRCA mutated, I think it’s great to have 2 options.
DR LOVE: Absolutely. I mean these drugs are not always a walk in the park. You’re absolutely right. It’s great to have another one. You never know, sometimes you have a side effect, and you don’t see it in the other one. Totally. Yeah, for sure. Interesting.
DR SAAD: Because I’ve worked with both of them in monotherapy, and they’re both very effective. I think it’s really the question of how we’re able to combine —
DR LOVE: Right.
DR SAAD: — what dose we’re able to use. But I’ve had great success with both of these drugs, and the more options we have the better it is for our patients and for our clinicians that have to work in the clinic. I think we’d all be disappointed if we only had 1 novel hormonal agent, right?
DR LOVE: But yeah. I wonder if it was just me, and I missed this, or maybe it wasn’t discussed as much in that original presentation a couple months ago in GU. But I didn’t appreciate the thing about the interaction with the abiraterone at all. I didn’t get it.
DR SAAD: So in terms of safety data, I think it’s safe, but patients have to be followed closely. What I presented are the results of 2 Phase III studies in combination in first-line mCRPC, but the field is really very active, so multiple therapies being investigated in the first-line mCRPC setting, so we’re anxiously awaiting TALAPRO-2. CASPAR hopefully will report in the future, in the near future. And we’re actually looking at mHSPC, where combination therapy in patients that harbor mutations as early as the hormone-sensitive setting is being investigated to look at the benefit, but also the risk of introducing these PARP inhibitors earlier since everything that we do earlier has a cost, financial, but also importantly risks in terms of adverse events long term. So all of these details need to be worked out with these trials that are ongoing, and we really look forward to a lot of new data.
So I’d like to just conclude that survival of men with mCRPC in the real world remains a huge problem. We still need to improve. There are really good first-line options, but early resistance and progression is a challenge. Second-line options are available, but many patients in the real world don’t get more than 1 line of effective therapy. And less than half the men with prostate cancer will receive chemotherapy before dying of prostate cancer.
So we really have to address that unmet need of these patients that either refuse or are judged to be not well enough for chemotherapy. So we really need to build effective first-line options for mCRPC, and PARP inhibition with a combination of NHT, in my opinion, fulfills an unmet need of effective and tolerable first-line combinations and hopefully will be the first combinations that we’ll be able to offer our patients. The question is will we offer it to all patients, like we saw in PROpel, or only in HRR-mutated patients. This is something that still needs to be addressed and hopefully the people who decide what we can offer patients will expand beyond HRR-mutated alone.
DR LOVE: I just want to ask you a general question, which is do you think that this strategy should be implemented in clinical practice.
There’s a medical oncology thing that I think really maybe started with breast cancer about the idea of maybe if you have a palliative situation, where you can’t cure somebody, in the long run a patient’s going to be better to get sequential single-agent treatments than combinations, even though the combinations are going to give you earlier — better relapse-free survival. That maybe you need to wait and see if there’s survival benefit before trying this into practice. I’m sure that’s been brought up many times to you in the last few months. What are your thoughts?
DR SAAD: Absolutely, and those are all fair comments. And when I say that we’ll hopefully have this — these combinations approved it’s never to indicate that I would give it to every patient who walks into the clinic. I think we have to look at patients individually, and some patients we know are not going to do well with single-agent abiraterone or enzalutamide or whatever we’re doing. And so when we can identify those patients, giving them the opportunity at least to have a more aggressive approach up front and maintaining stability for as long as possible. Because every time we go sequential in prostate cancer whatever we give in the second or third line is much less effective, if at all effective, and that’s been very disappointing.
So when we talk about younger patients, when we talk about patients with visceral metastases, when we talk about patients who have had chemotherapy in the hormone-sensitive setting, those patients do poorly with first-line NHT, really. And just looking at the results from PROpel the combination blows abiraterone out of the water in terms of how long it takes to progress. Obviously, with 800 patients we don’t have the power for survival of every subgroup, but I think we need to go beyond overall survival every time we try to introduce a new therapeutic option for prostate cancer or we’re going to be stuck with what we have forever. We really need to offer patients more than what we’re doing today.
So suggesting that every patient — obviously patients with mutations are going to benefit, and there I have no doubt that earlier introduction, up-front combination, is the way to go. Because we see the patients in PROfound, they do well, but it’s very modest improvement in overall survival and progression-free survival.
DR LOVE: So I never know whether to apologize when I bring up other cancers because there’s — PARP is a big part, obviously, of breast and ovarian and even pancreatic cancer. But I know that docs who take care of those cancers are asking these questions themselves. And one of the things they hear, particularly in ovary, but a little bit in breast, is this issue of platinums being effective, particularly in patients with BRCA germline or even somatic mutations. And in ovary they have this strategy they give platinum-based chemo then they follow it with PARP maintenance. Then they come over to prostate cancer, it’s a completely different model.
But is there any role for platinums, particularly in patients with BRCA?
DR SAAD: It’s a really good question, and I think that has been an understudied situation. Now our patients are maybe quite a bit older than the average patient in the other trials in breast and ovary, and platinums are being used in prostate, but the toxicity of platinums in the average patient with prostate cancer and the response rates have been disappointing. And the question of specifically BRCA-mutated patients is still an unknown and probably needs to be studied because I guess that is a poor man’s alternative. But the data is really supportive, at least with the PARP inhibitors that we’ve been working on for several years now.
But the question remains if you don’t have access, if you’re unable to get a PARP inhibitor, pay for a PARP inhibitor, I think in a BRCA-mutated patient platinums are a reasonable option, and it would be great if we could have a registry to try to get a sense of how these patients respond and tolerate. But personal experience with platinums have been difficult in our patients with prostate cancer.
DR LOVE: So this first patient seems like he kind of went through the path, let’s call it the old way in terms of PARP inhibitor.
DR SAAD: Yup.
DR LOVE: So you have a patient with BRCA2 germline mutation. He had a family history?
DR SAAD: Yes. Yes. Family history, daughter with breast cancer, and so the family was known to be a carrier. He had metastatic prostate cancer at diagnosis, but at the time there was not much interest in doing anything more than standard ADT, obviously, and there was really no reason to do more than ADT.
He became mCRPC 2 years following the start of ADT, was put on enzalutamide, and the PSA responded reasonably well, going down from 16 to 6. But the PSA rose about 15 months later, with radiographic progression with metastases in the bone and in the retroperitoneal lymph nodes. The patient came in for PROfound. He was referred for PROfound, and he went onto the olaparib arm and became a bit of our posterchild because he went on TV to talk about the time that the PROfound data was published in the New England Journal.
He tolerated the drug very well; minor fatigue. The PSA actually went to below 0.2, undetectable. Everything basically almost disappeared on imaging. Lymph nodes were all below 1 cm, and there was reduced intensity on bone scan, and he remained without progression for 18 months. So really outstanding results in this patient that really didn’t want chemotherapy.
DR LOVE: Before you continue, I’ve been asking this poll question to audiences, but I’m not sure I exactly know the answer, which is how do men with metastatic prostate cancer to the bone, like him, do in terms of cytopenias with PARP inhibitors compared, again, to other younger patients without bone mets. I hear people say that men do well, and I guess this man didn’t have a big problem with anemia or cytopenias. So is that the case? That even in fact when they have a lot of bone mets they still do pretty well in terms of cytopenias?
DR SAAD: That’s been our experience. And almost all the men we’ve given PARP inhibitors to have had bone metastases. And I think it’s actually more related to the line of therapy than to the extent of metastases. So the patients we’ve had the most issues with in terms of cytopenia, anemia —
DR LOVE: Interesting.
DR SAAD: — patients who were in the third-/fourth-line setting, so were exposed to chemotherapy. And this is why, at least in our center, this is what we would do now. So a patient who has a known mutation, who has failed an NHT, we will go straight to a PARP inhibitor rather than have them go through chemotherapy and then a PARP inhibitor. Because we feel that they’re able to tolerate it better and the earlier introduction seems to give better long-term results than when we introduce it post chemotherapy. And that was what was nice with PROfound is that we had the option to take patients that had been exposed to chemo or not.
DR LOVE: So it’s interesting that he ended up on lutetium. You were talking about quintuplet therapy, like multiple myeloma, I wonder if the next round is going to be bring in lutetium plus PARP plus abiraterone. But anyhow, interesting. I don’t know if you know Ross Camidge. He’s a lung investigator at University of Colorado. He was telling me about this idea of putting patients in a time capsule. Like you give them therapy that keeps them alive for a few years, and then allows them to get a new treatment, and that’s kind of what happened to him. He got lutetium —he lived long enough to get lutetium.
Anyhow, are they looking at — I know they’re looking at lutetium earlier, but are there trials looking at PARP, abi, and lutetium?
DR SAAD: Well, there are trials looking at lutetium in combination with PARP inhibitors, and it’s still early in terms of getting a sense for tolerability and all the rest. But clearly there’s a lot of excitement. They both have a mechanism of action that potentially would be synergistic, and so it makes a lot of sense to be looking at the combination. Obviously, in the VISION trial there was no patient who got the combination, so we don’t know yet, but things are looking good to be combining it. Obviously, lutetium in an earlier disease setting is being looked at in the earlier mCRPC and even mHSPC setting.
So this patient obviously progressed eventually and got what he was able to tolerate between docetaxel, but unfortunately started to progress after 8 cycles, in terms of symptoms and imaging. And we have an early access program to lutetium that has to go through the federal government, and so he’s presently stable after 2 cycles. So for now he’s still alive, very happy with what he was able to get, but realistic that we’ve still got work to do in terms of keeping these patients alive.
DR LOVE: This next case sounds really wild. The patient sounds like they’re in terrible, terrible shape.
DR SAAD: This was a patient who was sent really out of desperation from outside the province. I mean he came from quite a distance, from another university hospital that had nothing to offer. He had gotten basically everything for mCRPC, had chest tubes in for pleural effusion. He was ECOG 2, but able to come in and really wanting to get whatever was possible. His PSA was over 1,000.
He was eligible for the GALAHAD trial and was found to have a mutation of BRCA2. He started niraparib monotherapy at full dose, and after 1 month he started — we thought this was — this was — we couldn’t believe that after a month he was feeling so much better. His breathing improved, and after 4 months the tubes were — just no longer had anything coming out so they were able to be removed. He was swimming. He was going on vacation. His PSA declined to 1,200, and he resumed all his previous activities and continued to come in every month. He never missed an appointment, and he remained responsive for 22 months, and another posterchild.
DR LOVE: Wow.
DR SAAD: That even with a PSA that went above 5,000 he was still well enough. But obviously when he stopped responding there was nothing else that could be done for this patient but supportive care. But he thanked everybody for the opportunity for being part of a clinical trial that was really his last ditch and gave him more than an extra 2 years of life.
This next case is really one of the Phase I combination trials with niraparib. Because I led the Phase I combinations with niraparib looking at the combination of niraparib plus abiraterone, niraparib plus apalutamide.
So this patient had mCRPC. He had received docetaxel and then enzalutamide. A young patient. He was a 53-years-old patient, and he was referred for a clinical trial to see what we could offer this patient who had already failed enzalutamide and docetaxel, with rapid progression. He was in pain, requiring narcotics.
So he went onto open-label Phase I of niraparib and abiraterone to try to determine the dose, exactly, for the MAGNITUDE study. And this was published. So the full-dose niraparib and abiraterone were started, and the patient had some relief of pain, was having some symptomatic relief, was very happy about that.
Unfortunately, Grade 3 anemia and thrombocytopenia forced us to hold the niraparib, give him transfusions, and this allowed the anemia and thrombocytopenia to return to Grade 1. He resumed niraparib at dose reduced 200 mg BID. He tolerated the drug but again experienced some Grade 3 thrombocytopenia, and unfortunately the treatment had to be stopped. There was little success with cabazitaxel.
Again, this patient, widespread metastatic disease after chemotherapy made it such that it was hard for him to tolerate the combination that we tried to give him.
DR LOVE: Again, about how big was he? How much did he weigh, roughly? Was he a big man or small?
DR SAAD: He was a small man. He weighed about 65 kg.
DR LOVE: So again, thinking about ovary, that would push the dose down just to start with. So maybe he really needed 100.
DR SAAD: Yeah. I’m just wondering if at 100 we would have had much in terms of efficacy. Obviously this was a very advanced third-line setting, so I think there are multiple reasons —
DR LOVE: — probably his marrow wasn’t that great either.
DR SAAD: That’s right. Right. I think his marrow wasn’t that great, and this was all we were able to offer him because he was — there was nothing else that was possible for this patient than this Phase I study. And that’s the reality of Phase I studies. We’re trying to look at tolerability, PKs, and we always hope to see something in terms of efficacy.
DR LOVE: So yeah, it looks like this patient was on PROpel, right?
DR SAAD: Yeah. This patient was on PROpel. So this a 61-year-old patient who presented with moderate symptoms. He had no relevant past medical history, and I’m still following this patient. His PSA was 132. He had metastatic disease at presentation, both in the bones and lymph nodes and lung metastases. He had received 6 cycles of docetaxel. He had a relatively good response, with a nadir of 0.9 after 12 months.
We always hope to get undetectable PSA, but he went down significantly. But unfortunately it started to rise again 6 months later and then clearly became mCRPC 6 months after that with a PSA of 3.4 and started to have some symptoms related to his bone mets, and he had progression of lymph nodes. But otherwise imaging was stable.
So this patient was offered PROpel, which was open, of abiraterone plus or minus olaparib, in February 2019. He experienced some slight fatigue and a decline of hemoglobin from 13.3 to 10.5. He didn’t need any supportive care or transfusion, obviously. And 8 months later his hemoglobin was rising and now remains stable. But after 6 months of therapy his PSA was down to undetectable, and on imaging he’s in complete response. There is no measurable lesion left, there’s no change on bone scan, and he’s continuing to be on PROpel, seen monthly, remains in complete response. And so this is one of our patients, among several, that are doing very, very well on the combination. And he keeps coming back from Portugal with bottles of port, so I’m very happy to continue following this patient.
DR LOVE: That’s — let me ask you something.
Do you think that his clinical course would be compatible with a person who got abiraterone alone?
DR SAAD: There’s always that possibility, but unlikely. (1) The hemoglobin went down. And in a patient to stay this responsive for this amount of time is rarely seen in the post-docetaxel setting. The median time to progression for patients who had received docetaxel in the hormone-sensitive setting was in the range of about 13 months in the PROpel study.