Joyce A. O’Shaughnessy, MD: Not likely we’re going to be doing a whole lot of single-agent checkpoints, do you think, Aditya? What’s coming combination-wise with chemotherapy?

Aditya Bardia, MD, MPH: Yes. At this time, we don’t have a good predictive biomarker that can tell who would benefit with single agents. I think the rationale is, can we combine it with chemotherapy agents, which we know works? And maybe if we add immunotherapies into chemotherapy agents, it’ll have higher efficacy. So, I think the 2 key issues are, first, is there preclinical rationale for synergy? And I think the strongest preclinical rationale is with DNA-damaging agents, platinums or anthracyclines. Those are agents that can increase the expression of antigens, maybe even neoantigens, which can then make a tumor more primed to an immunotherapy approach. So, that is one thing to consider.

And the second is the toxicity. Is there any overlapping toxicity between an immunotherapy agent plus chemotherapy? The trials that I’m aware of that have looked at immunotherapy plus chemotherapy—there are several that are ongoing. But, at San Antonio, we heard the results of eribulin with pembrolizumab. And, in that trial, patients with metastatic triple-negative breast cancer who had received up to 2 prior lines of chemotherapy agents received eribulin in combination with pembrolizumab, and they showed an objective response rate of about 25% to 30%. So, there was some evidence of activity. Whether it’s better than eribulin as a single agent or whether a different partner would be better—I think these are all open questions. In terms of toxicity, there was really no major overlapping toxicity that was seen with the combination of these 2 agents. So, I think there is a role for a combination of chemotherapy plus immunotherapy. But, at the same time, we do need to investigate potential biomarkers for benefit for immunotherapy, particularly as a single agent.

And if you think of biomarkers in general for response to immunotherapy, I think there are 3 key components that we could consider. It’s all about the T cells. If there are no T cells, you’re not going to get a response. So, I think that’s why it makes sense that the presence of TILs is going to be associated with higher response to immunotherapy.

The second would be, how do we get the TILs? How do we get those T cells to come to the tumor? And in a large part, that’s the expression of the antigens. And that’s why tumors that have a higher mutation burden are more likely to express neoantigens. If you have more neoantigens, the T cells would come there, and they are the ones that have a higher response rate as seen in colon cancer with microsatellite stable tumors. And pembrolizumab is actually approved. You can use that assay, and it’s the first drug that was tumor-agnostically approved. It does not matter what subtype you have. If you have that phenotype, you could potentially use pembrolizumab as a single agent.

And the third thing in terms of biomarkers that I would consider is PD-1. I’m not sure that we know the exact cutoff, 1% versus 10%. But if a tumor really has a lot of PD-1 expression, those are strong breaks. And I think those would hamper the response to a PD-1 agent. And the data that I would cite are the use of pembrolizumab in lung cancer in a first-line setting where they compared pembrolizumab versus immunotherapy, but in tumors that had more than 50% of PD-1 expression. So, that’s a tumor that really has those brakes on. And, in that setting, they demonstrated a survival benefit with the pembrolizumab versus chemotherapy.

So, I think it’s quite complex. There are 3 different factors. The interplay between them, maybe you need a composite score. But I think we need to spend a lot of time and investment because we know when immunotherapy works, it really works well; it’s a home run. If you talk about home runs in the metastatic setting, that’s immunotherapy. So, as a community, I think we need to invest a lot of time and energy in really defining the biomarkers so we can benefit our patients.

Joyce A. O’Shaughnessy, MD: We have registration trials that have been completed, some are ongoing still, in the metastatic setting, to get these drugs to us in the hopefully pretty near future. Tiffany, the combination of nab-paclitaxel with atezolizumab, it had a very high response rate in a pretty small group. What’s happened with those data?

Tiffany Traina, MD: Yes. I think that we have to be careful attributing benefit to one or the other agents—it just is unclear. We have a good known cytotoxic chemotherapy that has activity. And so, I think it looks encouraging, certainly, but we don’t know what the incremental advantage is to adding in that oncology agent there. And there’s a lot of thought about whether nab-paclitaxel is the appropriate partner. Because, as you alluded to earlier, you can avoid the concurrent steroid use and the concerns about suppressing the immune system with steroid use. So, it seems like a rational partnership to move forward into further study.

Joyce A. O’Shaughnessy, MD: And I think we should hear first quarter this year, I think, about the IMpassion 130, which is nab-paclitaxel plus/minus atezolizumab, and let’s do a collective hope here that we will have that here fairly soon. I’m thinking we will, and it’ll probably be a sea change. Talk about what’s going to be your first-line treatment. You know what I mean? Again, though, we’re going to want, hopefully, some patient selection from that big trial to be able to understand, perhaps, who didn’t benefit. That would be helpful to know, for example.

Claudine Isaacs, MD: And that trial was fairly biomarker rich. So, again, we might get the initial signal without all of the different biomarkers, but, hopefully, some of those data will be forthcoming. Certainly, it was a large international trial that collected tissues and is a mandatory step for patients participating in the trial. So, hopefully, we’ll get some data on that.

Joyce A. O’Shaughnessy, MD: And then in Europe, and I guess the United States is participating, too, the paclitaxel plus/minus atezolizumab—that’s the IMpassion 131—that has closed now, I believe, as well. And then pembrolizumab, too, first-line setting for metastatic triple-negative disease. There is a chemotherapy of choice. It’s a taxane, or it’s a gemcitabine/carboplatin plus/minus the pembrolizumab first-line. I believe that’s still enrolling, but it must be getting close to finishing enrollment. So, I think we’re going to get a lot of data here pretty soon about that.

And one thing I’ll just mention is a very interesting paper in the JNCI earlier in 2017 by a group from Almac in Northern Ireland. And it looked at a preclinical study in triple-negative breast cancers, basically showing that triple-negative breast cancers that have homologous recombination deficiency, they obviously can’t repair double-strand breaks very well, which leads to fragmentation of the DNA. And what happens, I didn’t know, is that fragmented DNA gets into the cytoplasms, so you get cytosolic DNA, which incites an antiviral response, interestingly. And it’s a type 1 gamma interferon response. The cancer cell thinks it’s under viral attack. So, they have a 44-gene expression profile—not RTP7—a gene expression profile that can be done on paraffin that looks at this antiviral response.

And it initially was interesting. The initial work was done, published in JNCI by Kennedy and colleagues and Mulligan in 2014, using it as predictor for anthracycline benefit. It wasn’t bad at that, actually, the whole HR deficiency. But now they realize that the 44 genes all have to do with response to immunotherapy, PD-L1 expression, because what happens is that when the cells transcribe to 44 genes, and of course many hundreds of others, in response to a “viral attack” with a cytosolic DNA, it upregulates a number of chemokines, including CXCL5 and CCR10, which then bring in the TIL, here they come.

So, it’s a second mechanism beyond the double-strand breaks and the neoantigens being up there. In come the TILs because of chemokines, and up goes the PD-L1 expression. Loved it. Bring together the HR-deficiency story with the immunotherapy story, and there’s a really nice Cancer Discovery review article on this recently. Our goal is going to be, let’s get these triple-negatives into an HR-deficient state if they’re not already there. So, that’s really a really interesting area. Good data coming out in the next few months, we hope, with getting us our first approval here.

Transcript Edited for Clarity