Niraparib Maintenance for Recurrent Ovarian Cancer
Panelists: Bradley J. Monk, MD, University of Arizona and Creighton University; Robert L. Coleman, MD, MD Anderson Cancer Center; Gottfried E. Konecny, MD, University of California, Los Angeles; Katie Moore, MD, University of Oklahoma; Matthew A. Powell, MD, Washington University in St. Louis
Bradley J. Monk, MD: Let’s talk about niraparib, now the third PARP inhibitor to be approved based on a phase III trial published in the New England Journal of Medicine called NOVA. So, you have 3 platinum-sensitive maintenance trials: SOLO-2, which was just germline; now NOVA, which was germline and non-germline; and then coming is ARIEL3 with rucaparib. Matt, tell us about the NOVA trial.
Matthew Powell, MD: The NOVA trial was obviously quite exciting and, as you mentioned, it involved all those patients—203 patients—randomized in the germline BRCA cohort of a total 350 patients. So, you can see the breakdown of who had germline versus not. There were benefits in all the groups, as mentioned: benefits in the group that had germline BRCA, HRD-positive patients, and all-comers.
Bradley J. Monk, MD: Resulting in a broad label.
Matthew Powell, MD: A very broad label.
Bradley J. Monk, MD: No molecular testing required.
Matthew Powell, MD: Correct. Now, there is a biologic marker. These patients responded to platinum the first time, and responded to platinum the second time.
Bradley J. Monk, MD: That’s why I said molecular testing. I didn’t say there was no biomarker. The biomarker is getting response to platinum.
Matthew Powell, MD: And the tricky part is, there are patients who started on therapy 8 weeks after or within 8 weeks of that last platinum. The hard part for some our patients now is they met that biomarker a year ago, but the drug wasn’t approved. Can they get it now?
Gottfried E. Konecny, MD: Even if it’s approved, what’s clinically meaningful to you? I do think the paper is interesting that there are differences. The hazard ratio is significant in all subsets: BRCA-mutated, BRCA-nonmutated, HRD-positive. But for the HRD-negative, non-BRCA group, it was a P value that was significant. A 3-month difference in PFS, is that an endpoint where you would say, “I’m going to give someone a drug?” I don’t know.
Bradley J. Monk, MD: Let’s go through it. Let’s put some numbers together. So, in the germline BRCA population, where it should work the best, the hazard ratio is 0.27 or 0.28, whichever you look at, and 5.5 months to 21 months is the median PFS. You told us that SOLO-2 is 5.5 months to 19 months; it’s the same.
Gottfried E. Konecny, MD: Identical.
Bradley J. Monk, MD: It’s interesting that the placebo is exactly the same—let’s say similar, because nothing is the same.
Gottfried E. Konecny, MD: Yes.
Bradley J. Monk, MD: So, 0.27. In the wild-type, it’s 0.45—3.9 months versus 9.3 months. So again, pretty good, but the HRD makes it better. You got 0.27, 0.45, and the HRD is in between at 0.38. Then you say, “Well, it works in HRD negative.” So, it’s the other side of wild-type, it’s 0.58. You’ve got 0.58, 0.45, 0.38, and 0.27. It just gets better and better, depending on the molecular signature.
Gottfried E. Konecny, MD: I think what clinicians shouldn’t take home from this is that you don’t need an assay to select patients. I think this is only true for this highly selective group of exquisite platinum sensitivity, where the HRD assay may be not that accurate and it doesn’t capture all, and where you draw the cutoff for HRD. But if you go into resistant disease and the response rates for PARP inhibitors are between 20% and 30%, you need an assay to enrich for that. And so, I think we still have to keep pursuing an HRD score or whatever type of biomarker.
Matthew Powell, MD: One of the reassuring things was the PFS2 data that was presented that shows we didn’t really harm ourselves by giving that PARP now. They still responded to the next line of therapy quite well.
Bradley J. Monk, MD: The PFS of the next line of therapy is exactly the same—I need to use similar, nothing is exact.
Robert L. Coleman, MD: The one thing that I think we still need a discussion about—and I think Gottfried was getting to it—is in our practice, do we follow how this trial was conducted? And I don’t mean just this trial, I mean all 3 of them.
Bradley J. Monk, MD: What do you mean?
Robert L. Coleman, MD: In other words, if we give a patient platinum-based chemotherapy and get a good response to treatment, do we just stop after 6 cycles because it’s enough?
Bradley J. Monk, MD: So, at this meeting, there was a study. In the NOVA trial, half of the patients had PRs, and they reported at this meeting, “What about those PRs?” I would argue no, that before we didn’t stop, and that the study was done in a setting that I don’t generally practice. I don’t generally stop in patients who have a PR. But in the PR group, when they’re treated with niraparib in maintenance, it converts them to a response. The PFS curves are virtually identical to the CR. What it does is change the placebo. It’s better than the placebo, much better.
Matthew Powell, MD: Correct me if I’m wrong, these are PR patients who had a normal CA-125, correct?
Katie Moore, MD: They had a 90% reduction, less than 2 cm.
Matthew Powell, MD: And how well did placebo do at that same marker? The placebo also had that same effect to some level, right: 10% of placebo patients ended up having a CR as well, who were PRs?
Gottfried E. Konecny, MD: Yes.
Matthew Powell, MD: So, it gets to our imaging and measuring of disease, it’s a little bit interesting.
Robert L. Coleman, MD: Let’s break it down though. You’re saying it converts patients who have disease when they switch from a platinum-based combination to a PARP.
Bradley J. Monk, MD: No, compared to placebo.
Robert L. Coleman, MD: Exactly. So, to me, I completely expect that for patients who have disease at the completion of chemotherapy, if you gave them a different treatment that you know is working on all the other patients who we’ve given it to—olaparib, rucaparib, as we just mentioned—it would work. Of course, it’s therapy: it’s not maintenance.
Katie Moore, MD: But the study you need to do is PRs: continue on carboplatin or put them on niraparib, and then look at quality of life.
Robert L. Coleman, MD: Exactly, active control, not placebo.
Robert L. Coleman, MD: That’s my point, it’s active control.
Bradley J. Monk, MD: I think it’s a fair point.
Matthew Powell, MD: And what’s real disease? It’s hard to know. You have a few things showing up on a CT scan.
Robert L. Coleman, MD: So, the mistake in all of this interpretation is that we use the hazard ratio to support it, when the hazard ratio is dependent on the performance of the control arm.
Matthew Powell, MD: You compare to placebo; you look really good.
Robert L. Coleman, MD: You always look good, like Brad always says.
Bradley J. Monk, MD: If you can’t beat something, compete against nothing.
Robert L. Coleman, MD: Exactly.
Bradley J. Monk, MD: But we’re informed in clinical practice by FDA approvals because FDA approval allows reimbursement.
Robert L. Coleman, MD: Absolutely.
Bradley J. Monk, MD: So, my patient, she can always get carboplatin later.
Matthew Powell, MD: And it looks safe to do.
Bradley J. Monk, MD: And it looks safe to do with the same response.
Robert L. Coleman, MD: Same outcome.
Bradley J. Monk, MD: So, I’m going to give her chemotherapy now and when she responds—hopefully she will—I’m going to give her niraparib, and she’s going to have a good outcome. Then I’m going to give her carboplatin again, because that’s on the label.