Keytruda Added to Chemotherapy Improves Overall Survival for Metastatic PD-L1-Positive, Triple-Negative Breast Cancer

Jamie DePolo: Hello, thanks for listening. Dr. Hope Rugo is professor of medicine in the division of hematology and oncology at the University of California San Francisco Helen Diller Family Comprehensive Cancer Center, where she is also the director of breast oncology and clinical trials education. Dr. Rugo is also a member of the Breastcancer.org Professional Advisory Board. She is a principal investigator of a number of clinical trials looking at combining new, targeted, and immunotherapy medicines with standard treatments for both early-stage and advanced-stage breast cancer and has published hundreds of peer-reviewed papers.

At the European Society for Medical Oncology Congress 2021, Dr. Rugo presented final results from the KEYNOTE-355 trial, which was looking to see if Keytruda, also called pembrolizumab, and chemotherapy were better than chemotherapy alone as a first treatment for metastatic PD-L1-positive, triple-negative breast cancer. Earlier results from the same study found that adding Keytruda to chemotherapy improved progression-free survival, which is how long people live before the cancer grows for this type of breast cancer. These new results show that adding Keytruda improves overall survival, which is how long people
live whether or not the breast cancer grows.

Dr. Rugo, welcome to the podcast.

Dr. Hope Rugo: Thank you so much for having me.

Jamie DePolo: So, could you summarize the KEYNOTE-355 study for us, just so we're all aware?

Dr. Hope Rugo: Yes, and it's a little complicated, so, apologies for that because of the various concepts. But as you know, triple-negative breast cancer is defined by the absence of receptors. It tends to be — but not always, just the majority — very aggressive and rapidly growing. And when it metastasizes to other organs or sites, the median overall survival has
been less than 2 years. So, this is an area of tremendous unmet need where we've been searching for better treatments.

We actually have a new antibody-drug conjugate, sacituzumab govitecan, also known as Trodelvy, in patients who've already had treatment for metastatic disease, but the question about whether immunotherapy could benefit this subset of breast cancer was really important.

Why do we think that triple-negative breast cancer, in particular, would benefit from immunotherapy? Triple-negative disease has these immune markers. So, it's a little more aggressive. It has more differences from your own cells, so that your immune system tends to see it as more of a foreign object, and you need your immune system to be seeing the cancer for checkpoint inhibitors to work.

Also, triple-negative breast cancer has more infiltration of the tumor with the immune cells and has more of a marker, which is really important for these checkpoint inhibitors, called PD-L1. The checkpoint inhibitors target PD-1 or PD-L1, which are two factors that bind together. By preventing that binding, you allow the immune system of your own body to see the cancer.

So, the big question with this was — one is — would immunotherapy added to chemotherapy improve outcome for patients with metastatic triple-negative breast cancer, or metastatic TNBC. And then the second question is, does it only benefit patients whose tumors have this expression of PD-L1 or have PD-L1 in their tumor and immune cells?

So, those were the questions that KEYNOTE-355 was trying to answer. There were some other questions in there, which was that the other study that looked at a different checkpoint inhibitor, atezolizumab, allowed patients onto the study only if their cancer had come back at least a year out from their last chemotherapy for early-stage breast cancer. KEYNOTE-355 allowed patients to enroll who had recurrent disease at least 6 months out. Of course, patients who when they walked in the door had metastatic triple-negative breast cancer, called de novo disease, are also eligible.

Then, the other thing was that not all patients can receive Taxol, or paclitaxel. So, the study allowed three different chemotherapy regimens to be used: paclitaxel, or Taxol; nab-paclitaxel, or Abraxane; and then a combination called gemcitabine and carboplatin. Patients were randomized, who had metastatic triple-negative breast cancer who hadn't received any other treatment for metastatic disease, to receive either pembrolizumab or a placebo given intravenously every 3 weeks along with the chemotherapy that the patient and physician chose to give. So, you could only get that one chemotherapy, but it was a choice of three
different chemotherapies.

There was a plan in the trial to look at the patients based on their PD-L1 expression in the tumor and immune cells, as well as in the entire population, which is called the intent-to-treat population. The trial randomized 847 patients, and the randomization was 2 to 1. So, twice as
many patients received pembrolizumab versus placebo. And then the trial continued the pembrolizumab for 2 years. It's kind of an odd thing with pembro that that's what they do is they're continued for 2 years and then stopped, and patients continued the chemotherapy until their cancer grew or they had side effects.

So, what we showed was that in patients who had PD-L1-positive, metastatic, triple-negative breast cancer who received pembrolizumab, survival was longer than the patients who received placebo when both groups got chemotherapy. The median overall survival in the patients who had PD-L1-positive disease were shorter than we've seen before, we think because we included patients whose cancers recurred earlier than 12 months. So, it was just 16 months, but it increased to 23 months with the addition of pembrolizumab.

It's important when you think about median that you remember that that's just in the middle. So, that means that half of the patients lived longer than that, and one of the ways to look at that is to look at what we call a landmark analysis. When we looked at 24 months, 34% of the
patients were alive who had placebo and chemo, and 48.2% were alive who got pembro and chemo.

In the trial population, 38% of patients had PD-L1-positive disease. There's a special test that's associated with each of the checkpoint inhibitors. For pembro, or Keytruda, it's called CPS. It means the combined positive score, it uses a specific antibody, and it makes a scoring system based on how many of the cells, the immune cells and the tumor cells, have PD-L1 on the cell surface using an antibody. So, the definition of PD-L1 positivity for Keytruda is a CPS of 10 or greater, and that was 38% of the patients.

We also looked to see if there were different groups of patients who didn't seem to benefit as much, and the group of patients who had very early relapse from their cancer was a very small number. So, it's impossible to really know.

You heard earlier in the introduction that progression-free survival, or the length of time that the cancer is controlled before it starts growing again, was longer in patients receiving pembro by about 4.1 months. But we also looked at other factors, like how many of the
cancers shrunk more than by a certain percentage. We looked at how long the disease stayed controlled, and then we also looked at — which is something I think is always really important in patients whose cancers shrunk — how long did they stay decreased in size before they started growing again? We call that duration of response. And all of those factors were better when patients got pembrolizumab. And at 12 months, 56% of patients still had responding disease versus 38% when you compared pembrolizumab versus the placebo.

And of course, safety is a critical issue. We looked specifically at immune-mediated adverse events, or immune-related adverse events. The most common adverse events were related to the thyroid with either decreased or increased function of the thyroid that can be fairly easily
treated. And if you add together all of the thyroid disorders, it was about 20% of the patients. So, it's not uncommon, even though thyroid disorders are very common.

The other side effects were pretty uncommon, under 3%. Any organ can have immune inflammation from checkpoint inhibitors, but very low numbers, and if you recognize them early and start treatment, that works very well. No patient died from an immune-mediated adverse event or side effect.

So, this data actually led, along with the data in the early-stage setting, to final and formal U.S. [Food and Drug Administration] approval. So, it has regulatory approval in the United States and is now a standard of care for the patients who have PD-L1-positive, metastatic, triple-negative breast cancer.

Jamie DePolo: Thank you. Now, I do want to ask: Does Keytruda seem to offer benefits for people with PD-L1-negative metastatic breast cancer?

Dr. Hope Rugo: And that's a great question. You know, we first looked at that for progression-free survival, that first endpoint: how long does the cancer control? And we didn't see any significant improvement in the progression for survival in the patients who have PD-L1-negative disease, or even in the patients who had less PD-L1-positive disease.

So, you could have a CPS of 1. We haven't actually looked at the data if you just have between 1 and 9 for the score, and that data should be available by San Antonio. But we saw the same thing with survival. Although there's a numeric difference, it's small, and it's not
statistically significant in the patient population who have PD-L1-negative disease.

Jamie DePolo: Okay. Thank you. And help me to make sure that I'm understanding this
correctly. I've heard it explained before that PD-L1-positive cancers have more immune cells in them. So, the thought is that then when an immunotherapy medicine comes in, that particular cancer is more responsive because it's already got those immune cells inside of it. Is
that correct?

Dr. Hope Rugo: Well, it's not quite correct, but it's really close.

Jamie DePolo: Okay. Okay.

Dr. Hope Rugo: That's a good way of thinking of it. Actually, I think it's a really nice way of thinking of it!

But basically, there's two factors that we've identified as seeming to be important for outcome in triple-negative breast cancer and some other cancers, HER2-positive. One is PD-L1 expression on the cells. So, that means that there is a receptor, or a protein, on the surface of the cells, and that could be the immune cells or it could be the tumor cells. For example, atezolizumab, or Tecentriq, from the IMpassion130 trial, that test for PD-L1 only looks at PD-L1 expression on the immune cells.

And these are immune cells that are infiltrating, or invading, the tumor itself. So, that's the second factor, which are called TILs, or tumor-infiltrating lymphocytes. And there's been so much interest in that that there's actually a website maintained by colleagues internationally that actually tell you how to evaluate and enumerate, or detect, TILs.

And there's been really nice studies that have shown that the number of TILs in triple-negative breast cancer in early-stage disease is associated with how well the cancer does. Like, what is the risk of the cancer coming back with treatment. And even without treatment, in tiny cancers, the number of TILs make a difference.

So, there's two things. There's the expression of PD-L1, or the protein on the cell surface [of] tumor cells and immune cells, although probably immune cells are most important — we don't know. And then the TILs.

Jamie DePolo: Okay. So, you mentioned the test [Keytruda] uses to determine if a cancer's PD-L1-positive. Is the test that [Tecentriq] uses different? Could you talk a little bit about the differences between the tests?

Dr. Hope Rugo: Yes, and we just published a paper looking at a subset of patients in the Tecentriq trial, IMpassion130, and it took forever to get it published. So, I can tell that everybody is a little bit confused about these things, and there's a lot of detail, and pathologists who spend their lives thinking about this have very strong opinions.

So, the test for Tecentriq uses an antibody called SP142, and it looks at the percent of immune cells in the tumor bed that express PD-L1, and 1% or greater is positive. So, basically, it's like when you're testing for the estrogen receptor, and you use an antibody that has a little brown color, and the cells that are positive are brown.

So, it's the same kind of thing. And it's scored by pathologists, and so that's where having a 1% or not positive — there's a little wiggle room with all of these tests, but not a lot, because it's not a specific percentage, it's just any positivity is positive. It's easier to detect it.

Jamie DePolo: Okay.

Dr. Hope Rugo: In contrast, pembrolizumab, or Keytruda, uses CPS, which is a combined
positive score. And what it does is it uses a different antibody, 22C3, that they look at the positivity in the immune cells as well as the tumor cells, and then divide it by the number of tumor cells to get a score, and that's positive when the score is 10 or more.

What's fascinating is that the 1% or greater for Tecentriq is about 40% of patients. The CPS of 10 or greater for Keytruda is 38% of patients. But what we showed in that recent paper we struggled to publish is that the overlap is not complete. So, if you're positive for one, you're
likely to be positive for the other, but there are tumors that are positive for only one or the other test.

Jamie DePolo: Very interesting. Now, is this the first time an immunotherapy medicine has been shown to improve overall survival in breast cancer, specifically?

Dr. Hope Rugo: Again, great question. The first trial that showed the question of survival benefit was IMpassion130 with Tecentriq, or atezolizumab, and the chemotherapy drug nab-paclitaxel, or Abraxane. And the improvement was 7 months, which is almost the same as this trial, even though the populations are different. But this gets into, really, the weeds.

IMpassion130 had a hierarchical statistical design. So, you could only report the overall survival benefit in PD-L1-positive patients if the trial was positive in the overall, or intent-to-treat population, and it wasn't. So, it was inferred that survival was better. So, the way that's evaluated by the regulators is that you get accelerated approval, but you need confirmation with additional data to get formal, final approval. The KEYNOTE-355 had two primary endpoints, the progression-free survival and overall survival, too, but it didn't have that design, and you need different numbers and a different plan for the trial, and it really holds to the statistical design.

So, the problem is that they tried to have confirmatory data from another trial called IMpassion131 with Taxol, or paclitaxel, as the chemo partner, but that trial did not show a benefit, even in patients with PD-L1-positive disease, when we added atezolizumab, or Tecentriq, to Taxol. So, they didn't have the confirmatory data. There's no way to do another randomized trial based on KEYNOTE-355 with survival benefit. So, they withdrew the accelerated approval just a few weeks ago of atezolizumab and are waiting for a 1,500-patient trial in the early-stage setting that's already completed enrolling patients.

Jamie DePolo: Okay. I want to ask you, too, a little bit about progression-free survival versus overall survival. Because people I've talked to who've been diagnosed with breast cancer, sometimes they don't understand why is it so important for a medicine to improve overall survival as well as progression-free survival?

Dr. Hope Rugo: So, another excellent question, and one that we actually argue about as well, particularly when we think that, for example, if you did your trial in Russia, and patients did not have access to the other drugs, the survival might not be better — but that would just really be because of the treatment after progression on the study that you just didn't have access. A sad but true situation internationally that was brought up, actually, by the discussant of this presentation at ESMO just earlier today. And for that reason, there are sometimes situations where you might not see a survival benefit, but it is explained by access to subsequent therapy.

But in the majority of trials, the endpoints that we look at first — and I talked about some of the ones like duration of response — but the first one is progression-free survival and overall response.

So, why is that so important? Why don't we just look at response? So, if you have response, it means the tumor's shrinking to the treatment, and it's shrinking more to the experimental treatment than the standard. But the trouble is if that response only lasts for a few weeks — like the tumor shrinks, but 2 days later it's growing again — that didn't really help the patient. Just momentarily.

So, progression-free survival tries to get at that, where you want to see the number of patients whose cancers haven't grown by a certain amount during the trial, and compare that.

But then the problem is that you could generate resistance to the treatments by giving that treatment, right? So, what if you gave pembrolizumab, and the cancer was resistant to everything else afterwards because you changed something about the cancer? Well, then,
survival wouldn't be improved. In fact, survival could be even worse. What if you caused deaths from the treatment? Then survival also could be worse.

So, in a disease like this, overall survival is a critical endpoint. We always call it the gold standard, you know? I think it's a funny expression, but there you go. And I think that it is really important for us to see overall survival if we can.

Now, there are studies where overall survival is not seen, and that could be due to any number of reasons. And we have to look at them and see if the drug is still worthwhile, even if it doesn't prolong overall survival. Is it the patient population? The availability of other drugs? Changing treatments over time? All sorts of things.

I'll give you an example. For patients who have those germline — they're born with them — BRCA mutations, who can take the drugs called PARP inhibitors. The studies showed that the progression-free survival was really markedly improved compared to chemo, but overall survival was not. What we believe is that when you get PARP inhibitors, the cancer rapidly develops resistance, and in fact, sometimes can repair DNA, the problem that is being attacked by the PARP inhibitors. But if you treat maybe very early, in the early-stage setting before metastatic disease or even before patients have gotten some other treatment, we might see survival differences. So, that gives you an example where it can be a little muddy.

Jamie DePolo: Thank you for that. That was very helpful, because that's a question I've wondered about for quite a while.

Finally, I want to ask you, you talked a little bit about Tecentriq and how the breast cancer indication was pulled. So, what do these Keytruda results mean for people who've been diagnosed with metastatic, triple-negative breast cancer, PD-L1-positive, especially now in light Tecentriq is going to no longer be prescribed. My understanding is people who are on it, and the cancer's responding well, they can stay on it. But our website has heard from a number of people who are kind of panicked because their doctors are talking about switching them, even though they're responding. So, what do these Keytruda results mean, and sort of a secondary question, I am assuming that somebody, if they're on Tecentriq, it's just not possible to switch to Keytruda.

Dr. Hope Rugo: Well, I could answer the last question first. It is certainly possible to switch checkpoint inhibitors if your insurance covers it, and the insurance is going to be covering pembro and not atezo. But to go further from that — and we've been on Twitter about this, too. There is no reason — I mean, no, no, no reason — to stop atezolizumab, or Tecentriq, if you have disease that's being controlled. It should be continued, and the company, Roche, has really come out publicly, even in their press release, to say that they will work very hard to try and make sure that patients continue to have access who are responding to this drug. I have a patient who has been on single agent, off chemo for more than a year, and I would not take her off that drug. So although you could switch, I don't see a reason to do so.

What the approval means, which is great, and this data with survival, is that patients have an option, and that's really important. An option that improves progression-free and overall survival and is a new standard of care. I think we all wish that we still had both drugs available, but when we looked at this overlap in whether a tumor could be positive for both tests or one test, only very few patients had tumors that were only positive for the Tecentriq test. So, that's at least encouraging, that most patients will still be able to benefit from checkpoint inhibitors who have PD-L1-positive disease, and there is no reason why patients should not receive pembrolizumab, unless they have some major immune-related toxicity.

Jamie DePolo: Okay. Dr. Rugo, thank you so much. This has been hugely helpful in understanding this complicated research.

Dr. Hope Rugo: It's really been a pleasure to talk to you, and thank you for your interest in this data.