Case 3: The Use of Combination Therapy in Triple-Negative Breast Cancer

Hope Rugo, MD, FASCO: With that interesting data, let’s talk a bit about what we all do. Adam, do you test for PD-L1 expression in all patients who have metastatic triple-negative disease?

Adam Brufsky, MD, PhD: Yes, I do. And the assay does matter.

Hope Rugo, MD, FASCO: Ian, do you do that as well? Do you get a repeat biopsy, or do you use the archival tumor tissue?

Ian Krop, MD, PhD: Well again, as we talked about with the ER [estrogen receptor]–positive patients, it’s certainly important to get a biopsy to confirm the diagnosis. As you pointed out, the metastatic biopsy samples generally have lower levels of PD-L1 positivity. Whether it matters which one you measure from remains unclear.

We tend to measure off the metastatic sample, but a lot of times from a timeline standpoint, it’s easier to get the archival sample and get it done. I don’t think our institution has a specific policy. We want to get the results quickly so we can make a decision regarding their first-line therapy.

Hope Rugo, MD, FASCO: We had seen that initial data with checkpoint inhibitors as monotherapy. The data looked encouraging but then kind of became discouraging. Then we saw KEYNOTE-119. Adam, what’s your take on that? Does that mean pembrolizumab is not as effective, or is it more about the trial population with regard to the single-agent treatment?

Adam Brufsky, MD, PhD: It could be both. I think these are more heavily pretreated patients. When further analysis was done in KEYNOTE-119, it appeared that the patient with a CPS [combined positive score] of 20 seemed to actually have a benefit. That’s probably consistent with the patients who have a high tumor mutation burden.

It’s not a huge number of those patients, but those would be the only ones I would think about using pembrolizumab for in this setting, if I use it at all. Right now, I would not generally use immunotherapy after the first line, with the exception of these patients who have a high tumor mutation burden.

Hope Rugo, MD, FASCO: So you think the differences with KEYNOTE-119 had to do more with line of therapy than use of a single agent? Ian, do you share that idea? Is there any difference between agents that we know about? Is there something special about giving chemotherapy with immunotherapy?

Ian Krop, MD, PhD: I think those are 3 very good questions. In the metastatic setting, it does appear that being PD-L1 positive predicts benefit from the combination with chemotherapy, as well as single agent. As Adam pointed out, somebody who has a very high PD-L1 level may respond to checkpoint inhibitor monotherapy. But with the data from IMpassion130, and now with the hint that we’re hearing from the KEYNOTE-355 study, you’re seeing a benefit when added to chemotherapy. The overall outcomes in terms of response rate are much higher.

I think for most patients with triple-negative disease, you’re clearly more likely to see a benefit in terms of response when it’s done in combination with chemotherapy. So I think for most patients, and from a practical standpoint, it makes sense to use a combination. In terms of whether there is a difference between pembrolizumab and atezolizumab, obviously we don’t know for sure since we haven’t seen a comparative trial. And we probably won’t for who knows how long.

But now that we have the data from KEYNOTE-355, at least knowing that it met its endpoint, my guess is there probably is not a big difference in terms of outcome. They both seem like they are beneficial when added to chemotherapy. Again, as I said, I think for most people, that’s the right way to use them.

Hope Rugo, MD, FASCO: I agree. I think there’s something different about the responsiveness of our relatively immunologically cold disease that we specialize in when we give chemotherapy with checkpoint inhibitors. And there are, as Adam pointed out, some patients who may be uniquely sensitive to checkpoint inhibitors as single agents. Many of us who used those single agent trials have a single patient who’s far out on their checkpoint inhibitor, but it was only an N of 1.

Transcript edited for clarity.