Michael Untch, MD: I have just 1 question for my colleagues, because I see patients who have this devastating metastatic triple-negative disease. They undergo multiple lines of therapy for so long, and it happens that I do a lot of testing in their tumors, and I see that quite often, these really far-advanced diseases have androgen receptor positivity. At the moment in my institution, we don’t have an androgen receptor-blocking agent study.

Adam M. Brufsky, MD, PhD: Bicalutamide is generic; it’s not available.

Michael Untch, MD: Well, that is the reason I’m asking. What are you doing?

Michael Gnant, MD: We do this. We use bicalutamide.

Joyce A. O’Shaughnessy, MD: And because of the clinical cancer research paper, we can usually get it off label. I know Hope can sometimes even get enzalutamide. The data looked good there. But you have both enzalutamide and bicalutamide in the later-line populations. Generally speaking, they’re pretty strongly positive in AR, though it’s not the 10%. Usually, it’s higher. If you look at both studies, most of the patients who got clinical benefit had pretty high AR levels and low Ki-67 levels, generally speaking.

Adam M. Brufsky, MD, PhD: Have you been impressed with those data at all, really, for using androgen receptor blockers?

Hope S. Rugo, MD: The data is one thing. In practice, it’s another. I just haven’t had people have great responses. I know that there will be patients who sit there, but I also have patients who have that AR phenotype who’ve sat around on no therapy for years. And so, I think it’s totally bizarre, but there’s a lot of heterogeneity in this disease.

Michael Gnant, MD: However, similar to the other subtypes, we’re going to see more of this with better results by PARP inhibition and everything else. So, I think the field is going to expand. Androgen receptor blockades may be developed on treatment, and we struggle with testing. What is a mutation? What is the relevance of quantitative assessment? But I think there is more to come in this field.

Adam M. Brufsky, MD, PhD: I would agree. I think there’s something there. The last thing we want to end on here—and again, there are some abstracts of interest that are being presented at ESMO this year—is immunotherapy. It’s another big topic now for triple-negative breast cancer. Maybe we’ll start with Hope. What do you think of some of the immunotherapy data that are being presented now? And how do they fit in to where immunotherapy is going in triple-negative disease?

Hope S. Rugo, MD: It’s interesting. We thought breast cancer was not for immunogenic therapy, but these immune checkpoint inhibitors have taken the world by storm for many of the cancers that we didn’t have a lot of treatment options for. And so, we have to keep that in mind with breast cancer. It’s not as immunogenic. But clearly, triple-negative disease, maybe HER2-positive disease, and potentially the more proliferative hormone receptor-positive disease will be responsive to immunotherapy. We’ve obviously started with triple-negative disease, and now there are some studies going on in HER2-positive disease as well.

I think that what we’ve learned so far, and this goes along with preclinical data, is that immunotherapy is more likely to be successful if you treat patients earlier in their course of therapy. We saw lower response rates in later lines, and better response rates in the first line. And then, recently in the I-SPY trial with 4 doses of pembrolizumab, there was a suggestion of a clinically important improvement in pCR. Obviously, it’s a phase II trial, so it needs to be documented in larger studies that are actually ongoing already. That’s one thing we’ve learned, which I think is really important. These aren’t Hail Mary treatments because they don’t work as well in that situation.

We also have felt like the responses are not so dramatic that we want to push forward with single-agent therapy. The better responses in I-SPY, of course, were had when patients received pembrolizumab as the checkpoint inhibitor with chemotherapy. Now there are a bunch of randomized phase III trials looking at chemotherapy with checkpoint inhibitors versus chemotherapy as the standard alone, both in the neoadjuvant setting and in the metastatic setting, to really try and delineate that better.

And lastly, I think another area of really significant interest is giving a goosing agent. The idea is giving something that could be an immune agonist, whether that’s radiation therapy; chemotherapy; actual immune agonists, which have fascinating studies that are going on; PARP inhibitors; or a CDK4/6 inhibitor. All of those things are being studied.

Adam M. Brufsky, MD, PhD: Even with radiation therapy, there’s a trial here: the TONIC trial.

Hope S. Rugo, MD: The TONIC trial actually is a really fascinating idea with 50 patients, 10 in each arm—so tiny. They had no therapy with nivolumab, but radiation therapy, low-dose doxorubicin, oral cyclophosphamide, or low-dose IV cisplatin for 2 weeks. So that was the goosing approach. Let’s try to really push the immune infiltration. And then, they followed it with nivolumab. It’s actually very interesting. They had their median duration, a response of 10.9 months in the responders, and they had a total of 13 patients out of the 50 who had either a CR or a PR. So, we’re going to have to see what happens. That’s a tiny, thought-provoking trial.

We’re desperate, as we talked about with CDK4/6 inhibitors, for markers of response and resistance. In the trial looking at the first and later lines of therapy, we also looked at TILs (tumor-infiltrating lymphocytes). Sherene Loi and her colleagues did this in the samples that we collected from that study, which is a single-arm trial looking at only pembrolizumab. And that trial, which is also being presented at ESMO, suggests that the TIL level may predict, to some degree, in triple-negative breast cancer your responsiveness to checkpoint inhibitor therapy.

Adam M. Brufsky, MD, PhD: Didn’t we have TILs predictive for response in neoadjuvant therapy? Did we have a lot of trials?

Michael Untch, MD: We have, absolutely.

Adam M. Brufsky, MD, PhD: We had data from everybody.

Joyce A. O’Shaughnessy, MD: Triple negative is really, really predictive.

Adam M. Brufsky, MD, PhD: Right, TIL levels really predict for response.

Joyce A. O’Shaughnessy, MD: With the KEYNOTE-086 here, first-line metastatic disease in the PD-L1 positive group is 23%. But then, if you’re TIL level is high, it goes up to 39%.

Hope S. Rugo, MD: It’s a small number of patients, but yes. And those data will also be presented. There will be more data in San Antonio.

Adam M. Brufsky, MD, PhD: I’m interested, though, in the immunophenotyping of the TILs. I bet there’s going to be something there—maybe from next-generation sequencing or immunosequencing—that’s going to hopefully help us out, not just TILs, but something else.

Michael Gnant, MD: Yes, because there are still a lot of methodological issues around this. In general, it could be that they can be useful in breast cancer. In general, it is probably not very immunogenic but we can make use of immunogenics, and I would assume CD8 infiltration can be a reasonable marker for all the wonderful immunotherapeutics we have. Eventually, we can catch up with the other entities, because it is hurting breast oncologists that we are usually ahead of the pack, but now everybody else is carried away by these checkpoint inhibitors. But we are getting there.

Transcript Edited for Clarity