Emerging Concepts in Treating Triple-Negative Breast Cancer - Episode 12

Checkpoint Inhibitor Monotherapy for TNBC


Joyce A. O’Shaughnessy, MD: Speaking of the immuno-oncology, let’s transition to that because I don’t know about you guys, but I’ll predict that this 2018 is going to be the year that we get one approved for our patients, and they will be using them pretty soon, and I think it will be a sea change, truthfully, in triple-negative breast cancer. Let’s hope I’m right on that. Claudine, why don’t you just summarize the single-agent data that we have so far with the checkpoint inhibitors in the metastatic setting, and then we can maybe brainstorm a little bit. Are there any biomarkers we can look for in the clinic to help us?

Claudine Isaacs, MD: So, we’ve seen a couple of presentations this year, one at ASCO and then one at San Antonio looking at pembrolizumab. The KEYNOTE-086 trial had 2 different cohorts: one in first-line metastatic triple-negative and one in second-line or beyond. And in the first-line one, the objective response rate was about 23%. So, that was a pretty reasonable single-agent objective response rate. It was far lower in the patients who had received prior pretreatment. It was around 10% or so in terms of objective response rate, so that’s with the single-agent pembrolizumab.

So, again, it’s showing us a trend that we’ve seen in other trials as well, like in the nivolumab/olaparib trial, and it was also a much higher response rate earlier on with immunotherapy. In the atezolizumab trial, again, there was a similar response rate. It was also in the first-line setting. And that showed an objective response rate around 25% to 26%. So, again, getting the signals in about one-quarter of patients who are seeing benefits just with single-agent therapy in the first-line setting in particular, and then it looks like diminishing responses as you go further along.

Joyce A. O’Shaughnessy, MD: What’s of note to me on that is when they show you the spider plots and you see how long these people are going up. There’s a subset of them getting out there. You know they’re getting out there in a couple of years, which we hardly ever see in the metastatic setting. So, it’s really interesting, and those who have the CR or PR for atezolizumab or for pembrolizumab in the first-line setting have a really superior survival compared to those who progress or have stable disease.

So, definitely, like I said, signals are there for sure. And neither are approved for the metastatic population, but given the data, some of us are trying to get insurance coverage. If we cannot, there are access programs that we can utilize to get these agents for our patients. And I, for one, am doing some of that in selected patients. But the question is, in what patients? How do I select them?

Tiffany Traina, MD: Exactly, how do you select them? That is bit disheartening, as you said. So, in the first-line pembrolizumab study, the patients were enriched for PD-L1 expression. When you looked in the later lines where there was no enrichment and they compared responses to PD-L1 positive or PD-L1 negative, there was no difference, and it was single digits—I think 4% or 5% response.

Joyce A. O’Shaughnessy, MD: So, PD-L1 is not great.

Tiffany Traina, MD: That is not feeling like a good biomarker to help select our patients. And I know you probably speak to TILs as another possibility. I think we’re all looking for how to best select these patients besides just trying to use it early on and really trying to understand the optimal sequencing as well. I think your IMMU-132 data are really interesting to think that maybe whatever you use afterwards might potentially now have a primed tumor to respond better. I think it’s challenging to take those patients.

Joyce A. O’Shaughnessy, MD: Yes. And the TIL information, as far as I understand it—and jump in here—is that we do know for HER2-positive and for triple-negative in the preoperative setting, they are very nice data; that’s a higher level of TIL. There is an international consensus about how to read the stromal TIL, and with every 10% increase in the percentage of stromal TIL, you’ll get an improvement in disease-free survival in the early-stage setting. But we’re not using that, obviously, in the preoperative setting. And that was as a predictor for benefits in chemotherapy and from HER2-based therapy.

In the metastatic setting in the single-agent trials, there have been biopsies, of course, of metastatic lesions, going back and looking at those archival tissues and measuring the TIL as per the standard methodology that has been published many times, especially by Sherene Loi, who has really been the leader in this. And, interestingly, it is showing that there’s a difference in the percentage of TIL depending on the metastatic site. So, it’s the parenchymal lung and lymph nodes that have the highest percentage of TIL, and the chest wall is pretty devoid. You know, liver, bone, not much at all, which is interesting because in the original Rita Nanda trial, at the 18% response in heavily pretreated triple-negative patients with pembrolizumab, there was a small percentage of patients who went out about 2 years or so. And if you say, “What’s the phenotype? Who are those patients?” It’s the Isakoff patients. It’s the parenchymal lung, lymph node, and breast.

Claudine Isaacs, MD: We’re coining a few things here, today.

Joyce A. O’Shaughnessy, MD: And it’s all coming together. You know, the same ones who are very responsive to the DNA-damaging agents are the same ones who are really hyper, much more sensitive to the checkpoint inhibitors, which is a terrific lead in terms of how do you prove it as such. So, the thing, though, is we can’t really ask our pathologists to tell us about the percentage of TIL. Even though the methodology is published and has been standardized, it’s still something that the College of American Pathologists has not picked up, or ASCO hasn’t picked up, as one of our biomarkers such that there has really been standardization around the country and such. So, it’s not something, I don’t think, that’s really practical for us right now in terms of the selection. Because I think it comes down more than anything to phenotype and using it as early as possible as we can in the metastatic setting.

Transcript Edited for Clarity