Joyce A. O’Shaughnessy, MD: Let’s go on to the third data set, which is the KEYNOTE-522. We’re finally getting a big advance for patients with high-risk TNBC [triple-negative breast cancer] in the curative setting. Can you give us your thoughts on that trial, Sara?

Sara A. Hurvitz, MD: Absolutely. The KEYNOTE-522 trial was the study that led to the approval of the first immune checkpoint inhibitor for breast cancer in early stage disease. In this study design, patients who had triple-negative breast cancer that was either T1c with node-positive disease or T2 to 4 up to N2 were randomized in a 2:1 fashion to receive pembrolizumab every 3 weeks with chemotherapy, or placebo with chemotherapy. The chemotherapy was carboplatin given every 3 weeks or weekly, with paclitaxel given weekly, followed by AC [doxorubicin hydrochloride, cyclophosphamide] or EC [epirubicin cyclophosphamide]. Doxorubicin or epirubicin was utilized in this study, and the pembrolizumab was given throughout both the TC [docetaxel, cyclophosphamide] portion and the AC[doxorubicin hydrochloride, cyclophosphamide] or EC [epirubicin cyclophosphamide] portion.

Patients then went to surgery and completed pembrolizumab in the adjuvant phase or placebo. The primary end point of this study was pathologic complete response [pCR] in the intent-to-treat population and event-free survival. Interestingly, this study didn’t require PD-L1 expression for eligibility, but they did evaluate it. PD-L1 expression based on the CPS [combined positive score], which we talked about earlier, is important as a determinant for benefit in the metastatic setting. We haven’t seen whether PD-L1 expression is important in the early stage or curative setting for determining benefit from immune checkpoint inhibitors as far as breast cancer goes.

In the next slide, you can see that the primary end point at the interim analysis shows that the pathologic complete response rate is almost 14% better in patients who received pembrolizumab. Patients who achieve a pCR tend to have better outcomes. In this study, event-free survival is shown in the Kaplan-Meier curve, which was better in patients who received pembrolizumab. It’s a substantial difference with a nice hazard ratio, with close to a 40% improvement in outcomes and a nice separation of curves. This underscores the benefit we see.

We’re not showing it here, but it’s important that when you look at the outcome for patients who receive pembrolizumab based on whether they had a pathologic complete response rate, although patients who had residual disease overall did worse, it was a poor prognostic indicator. Patients with pCR and patients with residual disease both benefited from the use of pembrolizumab. Pembrolizumab appeared to work in both of those patients, including those who had residual disease after neoadjuvant pembrolizumab, where the difference was substantial.

With immune checkpoint inhibitors, we must be thinking about toxicity. I follow the study the way it was designed. I’m not using this for stage I triple-negative breast cancer, because although it’s uncommon, you can have serious autoimmune toxicities from immune checkpoint inhibitors, and some of them are permanent or even life-threatening. It’s important for us to keep in mind the risks associated with it as we decide who should utilize this, but these are exciting data. The outcomes are beneficial regardless of PD-L1 expression, which is why we don’t need to test patients’ tumors for PD-L1 in the early stage setting, which is in contrast to the late-stage setting, where we need to demonstrate that a patient’s CPS is 10 or greater to utilize pembrolizumab.

Joyce A. O’Shaughnessy, MD: Thanks, Sara. I agree. I’m not using this in stage I disease, just in stage II. Sometimes if we find a patient with T2 breast cancer—let’s say 2.53 cm clinically, no negative. We like seeing that Ki67 of 90%. We think the chemotherapy is going to work great. Do they need the preoperative pembrolizumab on top of it? I’m not withholding the pembrolizumab, because if you look at the study of stage II vs stage III, or the clinically no negative vs clinically no positive, there’s the same relative benefit. There’s still 6%, 7%, 8% improvement in IDFS [invasive disease-free survival] in stage II patients, clinically no negative. If they fit the criteria, which is clinical stage II or III breast cancer, then I’m doing it. Provided they don’t have a serious autoimmune preexisting condition, I’m utilizing. This is an unpredictable disease, and patients don’t do well with stage II/III TNBC. We need this advanced, big time. Sara, what do you do if the patient is a pCR or isn’t a pCR? What do you do in the adjuvant setting for these patients?

Sara A. Hurvitz, MD: If they’re a pCR, I complete the pembrolizumab. I follow the study, and if they’re not a pCR, I complete the pembrolizumab because patients appear to benefit from the it. There are questions raised by these results. Do you need pembrolizumab in a patient with pCR? Is it really benefiting them if they have a substantial response? We don’t know. All we know is that the study was designed to complete the full year, so I follow that study even in patients who have pCR, unless they’re having immune toxicity that would preclude them from safely doing so. For those who have residual disease, I follow the study. But then a question comes up: should we be adding something like capecitabine in the adjuvant setting? We have phase 1 data in the metastatic setting showing that you can safely combine capecitabine with pembrolizumab. You could do that, although it hasn’t been established in this trial, and capecitabine wasn’t part of this study design. Another question comes up: what if you have a patient with a BRCA1 mutation, and residual disease would you give a olaparib with pembrolizumab? It hasn’t been well studied, so it would be outside standard practice. It wasn’t utilized in this trial. It’s the art of medicine. I don’t have a good answer for that question. I’m interested in hearing what you’re doing in that particular situation.

Joyce A. O’Shaughnessy, MD: Hope Rugo did a Twitter poll on this. I don’t have the level 1 evidence, but as I recall it was a funny. It was a consensus among people that those patients didn’t do well. It was a big improvement, an 11% improvement, in IDFS to event-free survival with the pembrolizumab who didn’t get a pCR. But by 3 years, 33% of patients have recurred unacceptably high. I’m adding the capecitabine if they’re germline BRCA-negative, and I’m adding the olaparib if they’re germline BRCA1- or BRCA2-positive. I’m adding it because the safety is there. We don’t have the data in terms of efficacy, but in my mind, olaparib works perfectly well, and capecitabine ought to work equally well against that residual TNBC biology. Hopefully, the continuation of the pembrolizumab will also be of benefit to those patients, so I’m adding it. These patients have too much risk, so I’m going ahead and adding it.

Transcript edited for clarity.