Tiffany Traina, MD: So, I think still the most common histologic subtype is ductal histology and then perhaps lobular, which we rarely see overlap in TNBC. We do have an apocrine subtype, and as we’ll probably talk about a bit later, apocrine cancers often express the androgen receptor. And so, maybe we begin to think about some different biology there. Metaplastic cancers, which often classify as TNBC, we start to feel have perhaps a higher proliferation rate, which makes us maybe a little bit more concerned about the growth and behavior of that histology. But to echo Aditya’s point, I think it comes back to the clinical phenotype of the patient and their disease course. And if we do have biomarkers, whether they’re IHC biomarkers or nowadays even germline BRCA status, I would consider a really relevant biomarker as to treatment choice. I think that’s really trumping the molecular classifications, which are still in, I think, a research setting.

Joyce A. O’Shaughnessy, MD: Super, great point. I guess what I just heard is that when you’ve got metastatic disease, if you see that if this was an apocrine cancer, there may be a little connect-the-dots there to check for the AR in the metastatic setting. There will be clinical trials or possibly even, we’ll get to it later, some off-label use of bicalutamide or something. For metaplastic breast cancers in the metastatic setting, we might think about a checkpoint inhibitor. There are very preliminary data of possibly some benefit for the metaplastic cancers with checkpoint inhibitors. But it really is a good point when it comes to connecting a molecular classification with a therapy. We now have a PARP inhibitor, we now have olaparib—which we’ll get to in a little bit—but that really is an important point that is going to be 1 early, small subset but a distinction for some of the triple-negatives. Claudine, are there any phenotypes on the triple-negatives that are of particular interest? Do you treat them any differently based on clinical presentation?

Claudine Isaacs, MD: I think we think just the extent of visceral involvement might push us more towards one thing or another. There’s early evidence that some of the sites of disease might be more or less responsive to immunotherapy, and we can talk about that as well later. I think Tiffany’s point about BRCA is important, BRCA1 and BRCA2, because I think we now have, very recently, a therapy targeting that and that is FDA approved. And I guess the other way that I think of it is all of us are increasingly doing some molecular profiling on our patients, and we’re often doing it in triple-negative breast cancer because we don’t have the markers that we have for other cancers. And I think if we find something that is potentially targetable when we’re doing the molecular profile, for me, that allows me to think more about a certain trial versus another. That might push me to some of these basket trials that are out there—MATCH trial, or the TAPUR trial, or other trials—that really have specifically targeted therapies we might want to investigate in our patients.

Joyce A. O’Shaughnessy, MD: We’ll get to metastatic discussion there because that’s where sometimes some of the genotype, phenotype correlations are beginning to be at least hypothesized. But in early triple-negative breast cancer, we don’t really have a lot of therapeutic distinguishing. I guess with the possible exception of if you knew of a BRCA1 or BRCA2, you’d put them on the adjuvant OLYMPIA trial with a PARP inhibitor, adjuvant OLYMPIA. But for early breast cancer, subtyping is not really that useful, is it?

Claudine Isaacs, MD: I don’t think so, but I do think that the BRCA story is a little bit different. I think we’re all grappling where do we use a platinum in neoadjuvant or adjuvant therapy for triple-negative breast cancer. And my own practice has been that if I know that if somebody is a mutation carrier—and I know I don’t have perfect evidence for this—I would tend to give it to that patient. Unless they had a really tiny MRI-detected triple-negative breast cancer, I would think about integrating a platinum in their therapy today. And I recognize the data are not absolutely rock solid on that, but that’s a patient where I would feel that I would really want to try, to offer.

Joyce A. O’Shaughnessy, MD: Just on that point, if you knew, Claudine, that somebody was coming in with a PALB2, or RAD51D, or one of the others, has that family history structure, would you tend to lean towards a platinum in that patient, too?

Claudine Isaacs, MD: I probably would today. And, again, I’m saying something different today than I said 6 months ago and that I’ll probably say at a year, so I’m very cautious about what I say and how quickly these data are evolving. But, again, moving to the metastatic setting across different cancer types, we’re now starting to think about PARP inhibitors more in HRD-related tumors, tumors that have deficiency more broadly in HRD and starting to investigate. So, I think we would start to group those together more in the adjuvant and neoadjuvant setting as well. I think many of us, and I don’t know how much we’ll get into this, are really grappling even in all of our triple-negatives where do we include a platinum. I think if you have that as a clue, you can only do the best thing you can today with what you know today. So, if I had a patient who had a mutation in one of those genes, I’d think about it as well.

Joyce A. O’Shaughnessy, MD: I think what we’re saying is that really today, even though there’s marvelous science and hypotheses about how to subset triple-negative breast cancer, that we really don’t have anything clinically actionable, with the exception of the BRCA1/2 in terms of what’s really proven at the highest levels for practice, although there’s some interesting things coming along which we’re going to talk about more in depth. We really don’t have too much more. I will just say that what I think about a lot in practice, again, and this is really more of the metastatic settings, I do try to think about which patients may at that moment have the greatest possibility of benefitting from a platinum agent, who may have that homologous recombination deficiency—not that we have an assay today, we do not. But the Ki-67 is not, I don’t think, a bad poor man’s HRD surrogate. That’s my personal opinion. The faster they grow, the more likely they’re HR deficient. There are good data suggesting correlation between Ki-67 and gamma-H2AX with double-strand repair deficiency.

I think it’s important from a practical standpoint because if we look at Steven Isakoff’s first- and second-line metastatic triple-negative breast cancer trial—that’s the TBCRC-009 published in JCO in 2016—it was looking at cisplatin and carboplatin. And in the first-line setting, there was 11.5% of the cisplatin-based patients who had a multiyear disease control, multiyear sneaking up on cure because it’s going out 4 or 5 years. And then carboplatin was 5.5%, only in the first-line setting. No one in the second-line setting got that kind of very, very durable response. And I’ve cured by accident some people in practice with platinum, triple-negatives. I don’t know if you have. I have several. So, these are real data. What he has shown there, he could not find a biomarker for those patients. He could not. He worked really hard at it.

But I suspect where we’re going—I’m going to look into the future—we’re going to need to know at that moment not the primary breast cancer but at that moment to what extent does a patient’s cancer have homologous recombination deficiency. I think that’s shaping up but we’re not there. Though we can’t test for it, that’s shaping up as a really powerful potential predictor for DNA-damaging agents and possibly checkpoint therapy. That’s just sort of a look toward the future, but we really don’t have that possibility now to find those patients.

Transcript Edited for Clarity