Adam, say a few words for us about what therapeutic approaches you’re most excited about from an investigational standpoint that might take us into the future.

There are lots of things that people are doing now. They’re doing the usual of combining immune checkpoint inhibitors with PARP inhibitors. Someone on this call did review that—it may have been Kevin—about the CGAS pathway that’s activated by PARP inhibitors in immune cells and antigen-presenting cells, that may be involved in tumor immunity. You’re boosting them by relieving the suppression with a checkpoint inhibitor. It was a really nice theory behind it. A number of different checkpoint inhibitors are being used. Sacituzumab is used in earlier lines of therapy with or without checkpoint inhibitors.

Even trastuzumab deruxtecan is being used in the triple negative space. That’s not even thinking about all the other immune approaches that are in very early trials in the phase 1, phase 2 setting, bispecific antibodies trying to join various points of the immune system. People are trying to figure this out. The overall theme is that triple-negative breast cancer is an inflammatory disease. It’s the inflammatory or immune response that you really have to modulate or that is being modulated and being suppressed somehow as you get to later lines. How do we activate that? We call it making cold tumors, hot or whatever that is. That’s the other thing.

Of course, we haven’t even talk about all the AKT inhibitors. It turns out 10% to 15%, or more or less depending on whom you listen to, of triple-negative breast cancers have a mutation to the AKTp 10 and 4 pathway. Then there’s the whole PI3 kinase pathway. There’s this whole area there that could be exploited and is—capivasertib and ipatasertib are being used. A lot of things out there are exciting. The whole thing with sacituzumab and everything just jazzed everybody up about this field. At least it jazzed me up. When Suzette Delaloge or that group in France redoes this real-world analysis 4 or 5 years from now, we’re going to see some survival benefit in triple-negative breast cancer because of it.

Beautiful. Thank you so much. I’m going to give everybody a chance to go around. Let’s go around the Zoom here—give just 1 final thought. Anything you’d like to highlight from our discussion from triple-negative breast cancer, early stage or metastatic, from ESMO [European Society for Medical Oncology]? Let me start with Kevin. Do you have any last words?

 Yeah, as we’ve discussed, sacituzumab govitecan was really a huge story out of ESMO. As we’ve discussed, there is this story in terms of checkpoint inhibition in the metastatic, operable setting. That story evolves and continues. The other thing, in terms of future directions to think about, is what’s going to be the role of circulating markers in the operable setting like cell-free DNA. As prognostic and predictive, we think about some of these super-interesting agents in the metastatic setting, which may translate to the operable setting. Will that help us define risk and response better?

Thank you. That will be a brave new world, I must say. Sara Tolaney, how about you?

 First, I’ll echo Kevin’s comment that a lot of our discussion stemmed around at the beginning: Who needs preoperative immunotherapy? While we all hand wave and think about risk, I think we need a biomarker. I do hope we can learn to be able to tailor therapy for our patients better than we are. We need to improve outcomes, lessen the toxicity, make sure we’re not giving patients drugs that aren’t helping them.

I do hope we’ll learn more. I know there is a lot being invested to be able to discover potential predictors. Hopefully we’ll be able to get there. In terms of exciting combinations, there are just so many exciting things, which is very promising. Taking the sacituzumab story, a lot of us are really interested to better understand if ADCs [antibody-drug conjugates] work better with immunotherapy. Could it be a way for us to get those tumors that aren’t traditionally responding to immunotherapy to respond? For example, could you take a PD-L1–negative tumor and get it to respond to immunotherapy if it’s used to maybe a better backbone? An ADC could get us there. There is a lot of work being done to potentially allow immunotherapy to work in these colder tumors. We’ll see more to come.

Thank you. Such great thoughts, you guys. It’s really super. Sara Hurvitz, how about you?

 At the risk of being redundant, I will echo what my colleagues have said: Biomarkers are the key because triple-negative breast cancer is classified by what it is not. It’s clearly very old school, and we need to define subsets. Biomarkers are going to be the key. Not only genomics but proteomics are going to be very helpful in defining targets we can go after.

The second thing I would like to say is that with the Delaloge data that we’ve all been referencing, the rates of de novometastatic breast cancer are very high for 

-positive and hormone receptor–positive breast cancer. It’s underscoring how we are getting really good and sophisticated at how to treat in the curative setting. We need to get there for triple negative, so that the majority of the triple negative we’re seeing in the metastatic setting is de novo and not recurrent. That will move the survival forward.

Thank you. None of us can retire until we get there. We have to see that one too. Adam, you have the last word, with something you’d like to note from the top of your mind.

I’ll echo everything you say. We have a lot of new directions. As I think about this disease, in particular triple-negative disease, 1 of the things I’m going to try to do—because we have some pretty strong, hard-core, basic immunologists at my institution—is try to get them interested in this as a model system. They are at some of our other institutions. For the hard-core basic immunologist, the answer may rest in the tumor. But because the immune response has so much to do with the host that maybe there are biomarkers in the host that will tell us which way to go. I’ve started to interact with them. They do these 15-dimensional flow-cell sorting of immune cell subsets of the peripheral blood and single-cell RNA sequencing of these subsets after they flow them.

Maybe there’s something in there. It’s incredibly complicated stuff. It’s way over me, but it seems as if that may be where some of these answers may lie. It will also be in the tumor, it will be in circulating DNA, but the markers are there. I agree with everybody. We need markers for who is going to benefit from this. Because when you benefit, you do really well, and I think that’s fine. I’ll leave it at that.

Super. Thank you, guys. It’s so exciting talking to you. I am really struck by the energy people have to continue to push up that hill. It’s really apparent that you guys are asking the next questions. It’s really gratifying. Thank you for a wonderful conversation. Thank you to everyone who participated. I hope you found this really useful. Nice update on ESMO, with a lot of new data. We hope to be back after San Antonio Breast Cancer Symposium with another discussion. Thank you very much for participating.

Videos

Future Directions for Treatment of TNBC

Let’s spend a little time on toxicities in terms of hints for us all about how we manage toxicities. What do we have to lookout for, and how do we manage them? Sara Tolaney, could you say a few words about that?

 Sacituzumab is given on days 1 and 8 of a 21-day cycle. What we’ve seen is that the most frequent toxicity is really neutropenia. About 60% of patients do end up getting grade 3 or 4 neutropenia. Because of this, we did see in the ASCENT trial that almost 50% of people ended up using growth factor support. I very commonly will add on growth factor support when giving it. If they come in and their neutrophil counts are a little low, then I realize that they can’t really tolerate treatment at this dose unless they get a growth factor. I definitely introduce that early with any signs of neutropenia.

The other adverse effect to keep in mind is diarrhea, although I will say it is mostly low-grade diarrhea. The rate of grade 3/4 diarrhea is very low, and it was about 10% in the ASCENT trial. Most of my patients have intermittent loose stool, so maybe every once in a while they may need to take loperamide. I counsel them on that when starting the sacituzumab, but it’s not that they need prophylactic loperamide or anything like that for the majority of our patients. I also let people know that they will lose their hair. All my patients generally do lose their hair. Obviously, at this point in time many patients are going to go into it without hair if it’s in the later-line setting. For those with hair, it’s important to let them know.

I also generally premedicate patients, so on the day of treatment I usually give a Zofran and a Decadron prior to treatment. I have use Zofran at home as needed if they want. Some people will use something like Aloxi instead of Zofran, but either way, generally that’s what’s used. I don’t really have much trouble though with nausea. Honestly, most of my patients actually do pretty well with it outside the neutropenia and intermittent loose stool and fatigue. Those are the major things that we deal with.

It was interesting to think about. One of the reasons sacituzumab had a big impact in survival—those curves are very impressive: They really pull apart immediately and quite strongly continuing to pull apart; it’s an amazing set of curves—is that it really is possible to continue the therapy. It is something that we don’t usually have to stop. You don’t get into treatment limiting toxicity. I want to get Adam’s take on this because you and your practice have are very good stewards of resources. As Sara just said, about half the patients in ASCENT who are on sacituzumab end up getting growth factors.

On the 1 hand, you say, if they’re not having diarrhea or other toxicities and they just can’t get their day 8 on time, maybe we should just do a dose or 2 of filgrastim to make sure that they can get their day 8. The survival advantage started with the 10 mg/kg, so maybe we want to get the initial cycles and keep it up there. On the other hand, I forget what percentage of patients did require dose reduction, but there are of course always are patients who require dose reduction, yet you still see this very nice survival advantage in spite of the fact that some patients got a dose reduction. I have been going down to 7.5 mg/kg in patients who can’t get their day 8 and I want them to. I don’t want them to be missing day 8, so I’ve been going down.

On the other hand, 50% did get growth factors in this. That’s where we saw the survival. I know we don’t really know the answer about whether we just dose reduce without growth factors, because a lot of our guidelines suggest that unless there is definite, known impact on dose intensity in the metastatic setting, the dose reduction should be considered. What’s your take on that, Adam?

You just mentioned it, Joyce. Twenty-two percent of the patients had a dose reduction in the sacituzumab govitecan arm, so we have no idea what their PFS [progression-free survival] and OS [overall survival] were. That needs to be presented. The problem is that the company was bought by Gilead Sciences, Inc, and who knows what their priorities are going to be, but that data will probably need to be presented somewhere. That’s the first thing. The second thing is that 50% did require growth factors. Normally in a non-OS setting, we would try to avoid them, but maybe we need to rethink that.

Traditionally, the approach for my colleagues here and me has always been single palliative chemotherapy. You reduce the dose, and you don’t use growth factors, but I think I’m changing that. I’ve found as we do more of this, we’re using more growth factors. Initially we didn’t in a clinical trial. I use a little more dose reduction, but I’m starting to use growth factors more with this.

I must say, this is a remarkably non-cross-resistant agent. I’ve had a number of patients who really were quite primary refractory from the get-go and very symptomatic who have had responses. It does make me want to think more about what you just said about this has a survival impact here. At least in the beginning, I have to really try to keep the dose intensity up for these patients. I’m going to really rethink that.

I want to also mention the bystander effect about sacituzumab. Even if some cells don’t have Trop-2, maybe the expression is a little heterogeneous, when the SN-38 is cleaved—it’s a cleavable linker; it’s such a high concentration of the SN-38 on the antibody—you get diffusion to the other cells, and it can kill the cells even though they don’t have Trop-2. This really nice bystander effect probably helps play a role in the effectiveness of this agent. It’s a remarkable drug, and we’re going to see a lot of it in other cancers and it will be moved up into the curative setting very quickly.

Phase 3 ASCENT Trial in mTNBC: Managing Adverse Events

Let’s turn to the very important new data that came out of ESMO [European Society for Medical Oncology Congress], because we have a new survival impact for our metastatic triple negatives. Adam, you were an investigator in a sense, can you tell us about that?

 Well, we were all on the paper. The bottom line is that it’s uncommon in our business to actually have in the third line a disease that really was refractory to a lot of stuff in survival benefit. And this is pretty cool. I’ll go quickly because there’s a lot we want to talk about. The bottom line is ASCENT was 529 patients. Sacituzumab is a monoclonal antibody to the Trop-2 antigen, and it has a cleavable linker to SN-38. That is an irinotecan derivative, which is a Trop-1 inhibitor. It was 529 patients who were randomized 1:1 to sacituzumab, day 1 and 8 of every 21-day cycle vs physician’s treatment of choice. I think it was eribulin, capecitabine, or gemcitabine, and then there were some other ones, but the bottom line is it was physician’s choice of therapy.

Just to give you the top-line data, the prior anticancer regimens was a median of 4. I think that did include adjuvant therapy. They were the typical ones we thought of, but it was 4. They had at least 2 lines in the metastatic setting, probably 2 to 3 lines. There were some patients who had up to 17 lines, which is surprising. But the bottom line is the progression-free survival went from 1.7, and that’s really important. That 1.7 number is now the baseline. It’s the standard of care that we can use now when we start evaluating other agents in the third-line setting in triple-negative disease.

Everybody is progressing. It’s 1.7 from the time you get the next scan. But it was increased to 5.6, and that’s gratifying because that pretty much was what we saw in the phase 2 trial, so it replicated that. But the most important part of this was the overall survival benefit, which went from 6.7 months to 12.1 months. It had a hazard ratio of 0.5 in a previously intractable disease. There were better response rates, and the overall response was 82% vs 11% with the physician’s standard of care, which is what we all expect.

I don’t know if we’re supposed to talk about the adverse events that are predominantly hematologic and diarrhea. I think we will talk about that in a bit. But the bottom line is that the hazard ratio…went in favor of sacituzumab. This is pretty good data. It is pretty solid, and it really represents a standard of care. There is a reason that Gilead Sciences, Inc bought Immunomedics for $21 billion.

 That’s a good point. There you go. That speaks volumes.

 Obviously this is going to be expanded to the ER [estrogen receptor]–positive setting. I think we are all probably investigators on a TROPICS trial. But the bottom line is this: How often do we see this? It’s pretty dramatic. I wonder what the holes are in this study.

 I don’t know. Sara, are there any holes? I guess we’ve got accelerated approval, is that right? What do you think will happen going forward, and how do you incorporate sacituzumab? We’ve had it under accelerated approval. What do you think overall about this?

 I think it’s really exciting. The accelerated approval and Adiyta Bardia’s data were really quite nice. We saw the 35% or so objective response rate in heavily pretreated triple-negative breast cancer. We were so wanting some sort of positive data with a therapy like this in heavily pretreated triple-negative disease that we all leaped to use it without randomized data. I didn’t dream these data were going to be this good, and I was really excited when we saw that it was not just PFS [progression-free survival] but OS [overall survival]. But I do want to call out how horribly patients do with single-agent chemotherapy in the control arm. What we see in practice underscores how bad the outcome is in the second-, third-, fourth-, and fifth-line setting for triple-negative disease.

It’s great for patients. I’m incorporating it now in the third line setting and have been since the approval. I am tempted to use it in the second-line setting in the patient, especially for someone whose tumor is not PD-L1 positive or who has had a fast relapse. I’m trying to move it up a little, but this is an area where we’re going to see it move up, and we’re going to see different combination strategies as well. It highlights how important clinical trials are in this setting and in patients who are heavily pretreated.

 Thank you so much. Kevin, can you tell us a little more about kind of what sacituzumab is? Why was it so dramatically not cross resistant in this heavily pretreated population?

 Sacituzumab govitecan targets Trop-2, which is expressed on epithelial cells. Going back to the point that Adam made, we’re seeing benefits with this drug. Not just in triple-negative breast cancer, hormone receptor positive, 

 negative—those data have been published—but in urothelial cancer, lung cancer, etc. It’s a protein that can be broadly expressed. It’s linked with a stable linker to SN-38, the topoisomerase inhibitor. To the question that you had posed, in terms of cross-resistance, 1 of the things is that the payload is a different class of antineoplastic agents compared with taxanes, anthracyclines, etc. This subsequently has different toxicity profiles as well.

That’s part of why we’re seeing this nice activity a little later on. Just to reiterate 1 of the points that was previously raised, we’ve been in circumstances in which we’ve seen great data in phase 2 studies that unfortunately don’t always translate into phase 3 trials. The phase 2 data were published, and we’re also seeing a great response rate. We see this exact same response rate of about a third in this phase 3 trial. So it was nice to see these data in a randomized setting.

Phase 3 ASCENT Trial in mTNBC

 Before we go on to the new survival data with sacituzumab govitecan that came out of ESMO [European Society for Medical Oncology Congress], which I do want to spend some time on, let’s stop here for a second and summarize.

Let’s go around, so each of us can summarize how you in general approach the first- and second-line treatment in the very heterogeneous group of patients who are triple-negative breast cancer. Because we certainly have de novo, we have pretreated, we have rapidly recurring, we’ve got PD-L1 positive, we’ve got germline, 

, and we take a little bit of a different approach. It’s a tricky question. Kevin, let me start with you. How do you sequence these in the first- and second-line setting?

 I check PD-L1 status. If somebody is in that 40% or so range of having a PD-L1 positive tumor, I’ll give them a checkpoint inhibitor alone with chemotherapy. Even, for instance—I know we just spent some time talking about 

 mutations and the potential for PARP inhibitors—if I had a patient who was a germline 

 mutation carrier and had a PD-L1–positive tumor. I would tend to, outside a clinical trial, use the checkpoint inhibitor first because we can see some durability of response.

Although we can see some really nice responses with PARP inhibitors, the durability is not quite the same. Especially for those patients with checkpoint inhibitors who have long-term responses. There is some question in ongoing studies about combining those 2 agents, so we’ll see some of those data in the future. But I do think that in the absence of having a 

 mutation, this is somewhat dependent upon what they’ve received before: their time, their comorbidities, etc. The ASCENT data that we’re about to talk about were really important.

We keep talking about this disease-free interval that’s less than 12 months. If I had a patient who had an early relapse and then went on to get a checkpoint inhibitor, for instance, I would think about using sacituzumab govitecan in the second line. In the ASCENT trial, in patients who received therapy in the operable setting and reoccurred early, that did count as a line. I know we’ll spend some time talking about those data, but that’s generally how I tend to approach that in the first and second lines.

 Thank you. How about Adam? How do you approach first and second line?

 Like everybody else. Single-agent chemotherapy and sequential therapy still remain the standard of care: taxane, palbociclib, capecitabine, carboplatin-gemcitabine, whatever you use. That remains the standard, to be honest with you. If you’re 1 of those 40% of patients who are SP142 positive, it’s going to be CPS [combined positive score] of 10, probably at some point with pembrolizumab as well. They’ll get the checkpoint inhibitor plus probably a taxane, and a taxane will depend on which checkpoint inhibitor you use.

Second-line therapy at this point will almost always be single-agent chemotherapy, except for those scenarios in which you relapse pretty quickly, as Kevin said. In my case, that will be probably be eribulin. In the third line, as we’re about to talk about, everybody is going to use sacituzumab—at least in my practice, as many who can get it. Because most of us were on the paper as investigators, we know what the data showed. But the thing, that really is going to be pretty dramatic. That’s how I’m going to do it. We’re going to mix in other things as time goes on. But for me, that’s going to be the treatment approach.

 I agree with what they both said. The only additional point I want to make is that in patients who have both a 

 mutation and PD-L1 expression in the tumor, I would sequence first with the atezolizumab nab-paclitaxel. Given data that consistently show that a patient has the best chance of responding with frontline therapy, their immune system is more intact. That makes sense even if their disease-free interval was relatively short. I would still consider using something like pembrolizumab and gemcitabine-carboplatin as opposed to doing frontline PARP inhibitor because I want to get that immunotherapy in earlier.

 Thank you. Thanks for that key point. Sara, anything you’d like to add? Any other points on this?

 No, everyone did it really well. It’s just very critical to make sure we do get PD-L1 testing in our triple-negative patients, so you know in that first-line setting. As Sarah had mentioned, it seems that immunotherapy doesn’t work well outside the first line. We have very limited data in combination with chemotherapy outside the first line. But from the largest data sets, it seems it’s best up front. So I agree, getting PD-L1 status up front is critical to help make that first branching point. I also tend to use PARP after immunotherapy, even in my germline 

 carriers, because I worry about exposing them to immunotherapy later and not seeing that benefit. But we know PARP still has PFS [progression-free survival] benefit even in the later-line settings.

 I’ll add 1 thing. I’m really intrigued by KEYNOTE-119. We couldn’t really talk about it that much, but what Javier Cortés presented at San Antonio Breast Cancer Symposium last year that the CPS 20 patients—those uncommon patients in KEYNOTE-119—actually responded to single-agent pembrolizumab. I’m not sure what to make of it.

 Yeah, the higher the PD-L1 staining in KEYNOTE-119, the better the benefit. The more patients benefited more durably. That’s clearly a very powerful strategy for patients, so always checking it right from the beginning of the first line metastatic is key. Let’s not forget to test all of our triple negatives for germline 

. Or you can get NGS [next-generation sequencing]. If you get a somatic mutation, then go ahead and do germline to see if it’s in the germline. We have those PARP inhibitors for patients as well.

Sequencing Therapies for mTNBC

Let’s spend a little bit of time with a summary of where we are for patients with germline 

 mutations in the metastatic setting. Sara, could you give us a summary of where we are on that?

: We’ve all been excited to see the data that have emerged with PARP inhibitors for our patients with germline 

mutations. We have 2 phase 3 trials, OlympiAD as well as EMBRACA. OlympiAD had compared olaparib to standard-of-care physician’s choice chemotherapy, whereas EMBRACA had compared talazoparib in very similar design to choice of chemotherapy. Both trials did not allow platinum in the control arm and also did not allow patients who had progressed on prior platinum. Both studies very nicely demonstrated a statistically significant improvement in progression-free survival [PFS], both of them around a delta of about 3 months between the 2 arms favoring the PARP inhibitor over chemotherapy.

There were also nice analyses showing better quality of life in the PARP inhibitor arm compared to chemotherapy. We were all hoping to see larger differences, potentially, in terms of PFS, and also we were hoping to see a survival benefit, but neither study has demonstrated a statistically significant improvement in overall survival. There was this hint in the first-line subgroup analysis of OlympiAD that there was a trend toward better survival with olaparib compared to chemotherapy. That’s a subgroup analysis, and the overall studies did not show a survival benefit. Now, obviously it’s challenging because people do cross over.

In EMBRACA, about half the people on the chemotherapy arm crossed over to getting PARP, and some people go on to get platinum. It does make survival challenging to demonstrate. There are studies that are ongoing trying to look at adding agents to PARP to see if we can do better. We’ll certainly learn more as time goes on. It also sounds like we’re getting some new intriguing data that PARP inhibitors may also work outside of just germline 

patients. Nadine Tung, MD, had presented some data at ASCO [the American Society of Clinical Oncology annual meeting] that has shown that in patients who have somatic 

 mutation, so the mutation in the tumor rather than a germline mutation, had about a 50% response rate. So very similar to the response rates that we see in the germline 

Obviously this is a small group of patients. Some have reported maybe around 2% or 3% of our patients have somatic 

. Then there were also really impressive data in germline 

 suggesting that olaparib had about an 80% response. This is a tiny trial that had 11 patients, and I think 9 out of 11 patients had a partial response and the other 2 had stable disease. That’s a pretty big home run to have that level of activity. They are expanding that trial to get larger numbers of patients in the somatic 

 subgroups. We’ll get better numbers, but I think definitely there’s a hint that PARP inhibitors can have activity in those 2 groups as well.

Thank you. That was very exciting. Yes, we’ve enlarged the group of patients, I think, that can benefit. I think that was practice changing. Let’s look for those somatic 

mutations on our NGS [next-generation sequencing]. If you find germline 

, let’s use the Nadine Tung data to get preauthorization for a PARP inhibitor. I thought that was really practice changing. Thanks a lot, Sara.

Standard of Care for BRCA+ mTNBC

 How about you, Sara? What do you do with all the data at hand right now in your practice? What’s your approach to testing for PD-L1?

: Right now, we are testing with SP142 on that Ventana platform. Really, because as Sara Hurvitz very nicely reviewed that analysis that Hope Rugo, MD, had done, suggesting that the SP142 antibody was really the best predictor of benefit to atezolizumab. Given that, we did feel like we even had to validate the platform at Dana-Farber Cancer Institute in order to do that, because for the other diseases, our pathologists have been using a different antibody. We had to just do this for breast cancer, but they made it happen because of the strong data there are to support using it as a predictor for atezolizumab benefit. As Kevin alluded to, we are waiting to see if pembrolizumab is going to be approved based on KEYNOTE-355.

Obviously, as we discussed, that does utilize a different antibody with 22C3. We have had some discussions with our pathologists about what the impact of that would be at our institution. Are we going to start testing with one antibody vs the other, or both? They told us they will not test for both, and that we really have to make up our minds and pick one. I do think we are going to have to make a decision here because it’s hard to do all of these tests, and there is some discordance, even in Hope’s very good presentation. If you looked at the CPS [combined positive score] 10 subgroup and SP142-1, there was still some discordance, even though the CPS 10 did seem to have the most overlap compared to CPS 1 or 22C3 with 1. There was still some discordance. You could have one assay be positive and the other one negative.

This is why this field is so confusing. We’ll probably end up continuing with the SP142 since that’s what our institution has now figured out how to do. Then the question is going to be what do you pick? Are you going to pick using taxane with pembrolizumab vs nab-paclitaxel with atezolizumab? Are you going to use carboplatin-gemcitabine? It’s probably going to depend on the patient. The nice thing about KEYNOTE-355 is it did include patients who had early relapse. We don’t have that yet, we’ll eventually get that from IMpassion132, but we don’t have those data yet.

We’re stuck with a taxane backbone with atezolizumab, and I think that’s not what we want to give to someone who relapsed soon after adjuvant taxane, and we don’t yet have data for that. I probably would use carboplatin-gemcitabine-pembrolizumab in an early relapsed patient. That being said, if you look at the subgroup analyses in KEYNOTE-355, they didn’t quite do as well as the taxane group. We all have to be cautious because obviously the carboplatin-gemcitabine arm is more weighted with early relapsed patients who are bad performers compared to the taxane arm. That study is not powered to compare the chemotherapy. We cannot use it to say one is a better backbone than the other. It does make us wonder, obviously, how good is carboplatin-gemcitabine-pembrolizumab in an early relapsed patient? How much benefit are we really seeing in that specific subgroup?

Because we do see them, unfortunately, not infrequently. In that group I probably will use pembrolizumab. I will also throw my hands up in the air about IMpassion131. One thing I will point out, and I was totally surprised, I really thought it was definitely going to be true that paclitaxel-atezolizumab would be the same as nab-paclitaxel-atezolizumab. It is interesting if you look at the study design, the number of patients in the control arm who are PD-L1 positive is very small. That control arm is performing much better in terms of survival than the IMpassion130 control arm. It does make you wonder why is that? Is there some difference in the patient populations?

I think Kevin was alluding to that. Technically it’s the same number of patients who had had prior taxane exposure, but we don’t know the nuances of that. How long ago was that prior taxane? Was it different in the 2 studies? The number of de novo patients is also smaller in IMpassion131. There are some very subtle differences. We don’t know the differences in PD-L1 testing on primary vs metastatic, and there are some very small differences that potentially could pan out here. I agree I didn’t really think steroids were going to make a difference because we saw in the early disease setting it probably doesn’t matter. This is a different setting, it’s a metastatic population where the immune microenvironment isn’t as good.

It sounds like most people do continue their steroids through the whole time they’re on chemotherapy for most patients in IMpassion131. They were getting it for quite a long time, not just for a couple of doses, so maybe it had a little impact. I honestly don’t know what the right answer is.

: Can I make one comment before you go on?

: The steroid issue was an interesting one because if you look at the adverse events of special interest, there were less in IMpassion130.

: I agree. Maybe it’s just mismatching. I think it’s still the most likely explanation.It’s likely that PD-1 low positivity, it’s a patient selection issue, maybe a random patient selection. But I’m still looking at that steroids and looking at the fewer adverse events of special interest.

 Yes, that’s an interesting point there. It sounds like’s in summary, we’ve got 2 trials now with now with nab-paclitaxel and atezolizumab—the IMpassion130 with a survival advantage, that’s firm, and then the IMpassion031 in the preoperative setting that has a very nice delta on pathologic complete response. Atezolizumab is really doing well with nab-paclitaxel, that ought to be the standard of care for our patients who we’re going to give a taxane to, correct me if I’m wrong. There are some patients where the taxane may not make sense for a variety of reasons. Then we do have the gemcitabine, carboplatin, and pembrolizumab. It’s not FDA approved yet, but it’s not an unreasonable choice for patients where taxane is not the best choice for them.

Then we do have the pembrolizumab approval as well for tumor mutational burden high. I don’t know if that’s actually approved, but the MSI [microsatellite instability] high is approved. We have data in metastatic breast cancer that’s tumor mutational burden high, where pembrolizumab can have a positive impact as a single agent. There are some durable responses in those patients, and then MSI high. We don’t find those very often when we do next-generation sequencing, but we will find them occasionally. We do have the opportunity to use pembrolizumab for those patients.

We will stay tuned as some of the data mature, like KEYNOTE-355 on survival, and as we see other trials that we’ll talk about a little bit later to try to improve upon the impact with the immune therapy in the metastatic setting. Thank you guys so much for that summary. I think that brings us right up to date about where we are.

Testing for PD-L1 Expression in mTNBC

I want to come back and get everybody’s take on this. But let’s bring in KEYNOTE-355 before we have a larger conversation. Sarah Hurvitz, could you mention KEYNOTE-355? We saw those data at ASCO [the American Society of Clinical Oncology annual meeting], and could you talk a little bit about testing, like is any assay okay for PD-L1?

: The 355 clinical trial was also in the first-line setting, as Adam indicated. It had a little different disease free interval that was allowed. Patients could go on if their disease free interval was at least six months on this trial from treatment in the adjuvant setting, as opposed to IMpassion130 where it was 12 months. And they allowed treatment of physician’s choice based chemotherapy, so either nab-paclitaxel, plain paclitaxel, or gemcitabine-carboplatin. And interestingly, but not surprisingly, about 55 % of patients chose to use gemcitabine-carboplatin, which omits the need to lose one’s hair. They didn’t report how many got straight paclitaxel versus nab-paclitaxel, when I looked through the slides.

Maybe somebody else knows what the percentage was, but at any rate, patients received either the placebo or pembrolizumab, and they determined essentially the triple negative status as well as PD-L1 status. They used this thing called the CPS [combined positive score], which is a combined score looking at the number of PD-L1 positive tumor cells, lymphocytes, and macrophages divided by the number of tumor cells. Then they based their outcomes on this. When they looked at the progression-free survival [PFS], it was a hierarchical statistical design. First they looked at the PFS for CPS of 10 or greater, and the hazard ratio was 0.65, and it met their statistical significance. The median PFS was 9.7 months for pembrolizumab vs 5.6 months.

This was a nice 4-month, almost, improved median PFS. Then they go to the second step, which is looking at the median PFS in patients with a CPS of at least 1. Now, although there’s about a 2-month improvement in PFS, the hazard ratio isn’t as good, it’s 0.74. And while the 

value was .0014, it didn’t meet the predefined level of significance. They weren’t allowed to statistically look at the next question, which is the overall intent-to-treat population. They looked at it numerically. There was about a 2-month improvement in median PFS, but they wouldn’t give a significance value because they hadn’t met that step before. We see that there is this benefit with pembrolizumab. We don’t know if it’s dependent on which chemotherapy you choose. That wasn’t didn’t appear to be a factor. Some patients did get paclitaxel, and the benefit seems to be in the patients who are CPS 10 or greater.

Then the question is how do we test our patients and decide what therapy to give them? We’ve got the SP142 assay for PD-L1 in the tumor immune cells if we’re going to choose atezolizumab. Do we use the CPS score with the 22C3? This is just mind blowing. If you’re a general medical oncologist, how does one keep up with this? Well, Hope Rugo, MD, actually did a nice analysis and presented at ESMO [the European Society for Medical Oncology Congress] in 2019, looking at the different assays. And actually the SP142 antibody for PD-L1 expression, if you’re positive for SP142, then the other assays are likely to be positive, but the opposite can’t be said true.

If you’re looking just at the IMpassion130 data, and the analysis she did was somewhere around 600 patients where tumors were looked at by these 3 different antibodies: SP142, 22C3, and the SP263. She showed that the PFS was most strongly associated with the SP142, so it was most associated with benefit from atezolizumab. We don’t know that for other immune checkpoint inhibitors. This field is rapidly evolving and rapidly becoming quite confusing, in my opinion. We are going to continue to need to do tissue analyses to better understand how we need to test to identify patients who will benefit.

Thank you. That’s fabulous, great state-of-the-art summary, honestly. I want to ask Kevin and Sara Tolaney, given all the data we’ve just heard beautifully summarized, where are we right now in practice? We’ve got survival with IMpassion130, final survival data. Albeit, the intent-to-treat population was not positive for survival in IMpassion130, but the PD-L1 subset was. Negative, atezolizumab and paclitaxel, positive progression-free survival, but not yet survival. We don’t have the data yet for KEYNOTE-355, so we have a gemcitabine-carboplatin opportunity there for patients. How do you use this in practice? I’m also curious how you guys test in your institutions. Do you have multiple antibodies, do you tend to use 1 antibody? Let me start with Kevin.

: A lot of really great questions in there. As Sara Hurvitz had mentioned, the antibodies are not all the same. There are different pre-analytic and analytic considerations between the antibodies. To me, the specific biomarker associates with that specific agent. For SP142, that’s the predictor that I utilize for determination of whether to use atezolizumab. As Adam said, I also think we’re all a bit surprised, disappointed, and baffled by the results of IMpassion131. At this very moment, the approval really just is for atezolizumab in patients with PD-L1 positive tumors, but later in the year we’ll see whether this expands to pembrolizumab based upon KEYNOTE-355.

One of the distinguishing features, as was already mentioned, was the 12 vs 6 months and gives the flexibility of a different chemotherapy regimen. At the end of the day, to go back to one of the things that Adam had said, I wish I had a great explanation of why there is this difference between the 2 IMpassion trials. There has been some hand-waving, could it be the difference of the 2:1 randomization, and a smaller study, and then steroids for Taxol? But ultimately, as Sara Hurvitz previously mentioned, in the neoadjuvant setting patients got Taxol and they received steroids. We saw that that was a positive study, so I think that there is still much to learn.

I wonder, in terms of the differences between those 2 studies, whether there is just, by chance, a difference in terms of the biology of the tumors in those patients who were enrolled. There seems to be this subset of patients who are really driving a lot of the benefit. Is it possible that there were more immunomodulatory tumors in IMpassion130 compared to IMpassion131? Is there a difference in terms of the ratio of the subset of patients who are really long-term benefiting with those patients with durable responses? I don’t think we quite know a great explanation for the data that we have right now.

It sounds like in your practice, the atezolizumab, nab-paclitaxel is the standard of care for the SP142. Is that what you guys use, Kevin?

: Yes, that’s exactly right. For those patients who don’t have a recurrence within 12 months of their operable therapy, it’s to use SP142 and to use nab-paclitaxel along with atezolizumab, with the consideration of neuropathy and things that would concern me about using a taxane.

Would do ever consider the gemcitabine-carboplatin and pembrolizumab for people who recur more quickly or something, or who decline alopecia-induced chemotherapy?

: If this becomes an approved regimen, for sure. That’s one of the benefits of that study, the lower period from time to reoccurrence, as well as giving drugs that have different mechanisms of action, and patients may not have developed chemotherapy resistance.

KEYNOTE-355 Trial in mTNBC

Let’s transition to metastatic breast cancer because in triple negative there were new data that came out at ESMO [the European Society for Medical Oncology Virtual Congress], some of them very important, potentially practice changing. Let’s turn to Kevin, and he could give us an overview of the general strategies we have in metastatic triple negative breast cancer. Could you discuss overall outcomes and prognosis because we will get into how do we all sequence things in our own practice, but just from a general standpoint, Kevin?

: Sure. I think that if there is 1 subtype of breast cancer for which we really need additional therapeutics, it’s this subtype of breast cancer. Thankfully, we have had recent approvals of new drugs. In the frontline setting, for those patients who have PD-L1 positive tumors right now, there is the approval of giving atezolizumab, the PD-L1 inhibitor, in combination with nab-paclitaxel. We’ll talk about other studies that were recently reported with checkpoint inhibition in the metastatic setting. Also, for patients who have germline 

mutations, we have PARP inhibitors that are approved for patients with HER2-negative disease.

Then we’ll also talk about the approval of sacituzumab govitecan, which is an antibody-drug conjugate, targeting Trop-2, that’s linked to SN-38. For a long time, all we really had for these patients were chemotherapies, and now with the advent of a new antibody-drug conjugate as well as PARP inhibitors and checkpoint inhibitors, that’s really changed and is improving the outcomes of some of our patients.

Thank you. There’s been some progress, but unfortunately, median overall survival is still in the 16-month range. In France, Suzette Delaloge, MD, MSc, showed at ESMO this lovely real-world evidence database from across academic centers in France going back to 2008 through 2016/2017, that the median overall survival in metastatic triple negative disease is about 14 months. It hadn’t changed. It is very sad, so we definitely need improvement in this group of patients.

Now, that did not take into account any of the new checkpoint inhibitor data that we’ve seen with atezolizumab, the new approval, of course. That is hopefully going to make an impact. But let’s talk about checkpoint inhibitor therapy in the metastatic setting. Adam, could you update us on what we’ve learned from IMpassion130, IMpassion131, and how do you today use the checkpoint inhibitors in your metastatic practice?

Suzette Delaloge’s real-world analysis is really cool, and it says that we made huge progress in HER2-positive disease. We’re getting there. Actually, those data were accumulated before CDK4, so we’re going to get there with ER positive, and it will look pretty good. I think the stuff we’re going to present and talk about in the next half hour is going to tell us where we’re going to go, at least in the first stage with this, and probably help it. The first one was IMpassion130 that everybody’s familiar with. It was 902 patients with first-line metastatic advanced triple negative breast cancer with a treatment-free interval greater than 12 months.

This is really important, and I think it is the biggest difference between this trial and KEYNOTE-355 that we’re going to talk about. The patients in this trial all got nab-paclitaxel with or without atezolizumab and were treated until progression. The bottom line is that in the overall data, in the initial cut of this data, in survival and in PFS [progression-free survival], there really didn’t seem to be significant differences. They only appeared to occur when you looked at the PD-L1 subtype. You had to have, again by the SP142 assay, you had to be 1%. Basically, this is now the final analysis of these data, looking at it, actually the overall survival, overall as well as in the PD-L1 positive population.

In this trial again, I think the median follow-up was fairly long. There was a small increase in overall survival, in the intent-to-treat population, it was about 21 months with the checkpoint inhibitor vs 18.7 months. It was not statistically significant. The hazard ratio was about 0.87. When you look at the PD-L1 population, the difference was quite substantial. It was 25.4 months, which is pretty much what it had been in the past with some of the other cuts of the data, vs 17.9, about 18 months overall survival in the PD-L1 negative population. That was a hazard ratio of 0.67, which was statistically significant.

It was interesting. It was really the only group across all of the 4 subgroups. You had to have atezolizumab, nanoparticle paclitaxel, and be PD-L1 positive, and basically you had the benefit. The other 3 sub arms of the trial didn’t really. When you look at the hazard ratio for overall survival, the big groups had no prior taxane treatment and people who were PD-L1 positive. This shows us thatthe long-term survival is maintained in the study. That was IMpassion130, and I’m really curious to hear what my colleagues would have thought going into ESMO or before we heard the press release. We had seen the KEYNOTE-355 data with paclitaxel that looked like it was going to work, and we figured it was just like a throwaway trial. They just did paclitaxel and it was going to work, and it would be the same, and we’ll go on and move on.

The surprise, and David Miles, MD, presented this, it was pretty surprising in IMpassion131. The IMpassion131 was also a fairly substantial trial that randomized people to paclitaxel with or without atezolizumab, very similar design. There was really nothing different in this study, and it was basically a dead negative. I looked over these data, I pored over it, as I’m sure we all did, trying to find a subgroup. It was a little smaller. It was a 2:1 randomization for atezolizumab to placebo, but nonetheless there was no difference. The hazard ratio was 0.82. You look at these curves, maybe after 6 months there may have been a tiny separation for atezolizumab, but really not much.

It’s crazy that the forest plot was dead negative. There was little bit of response rate benefit. The overall investigator-assessed response rate in the PD-L1 positive population was 64% vs 55%. But really in the overall survival analysis, it was flat negative and if anything, the hazard ratio trended to touch toward the other group. Although when you really look at the data, the hazard ratio crossed 1, but there was a numerical difference. It was 28 months vs 22 with the paclitaxel. Is it simply play of chance? I don’t know, but there is no separation of these curves whatsoever. That’s the big mystery, and I’m really curious to hear what my colleagues have to say. With atezolizumab at least, we have benefit for survival in the PD-1/PD-L1 population in the first line with nanoparticle paclitaxel. But with non-nano, with standard paclitaxel, we don’t.

Yes, no question. Amazing. I think we were all taken aback, including in the PD-L1 positive population, in the intent to treat as well as PD-L1 positive, no benefit from atezolizumab in combination with regular paclitaxel.

Standard of Care for Metastatic TNBC

Just to finish the early-stage setting, I’m just going to turn us to Sara Tolaney. We definitely have an unmet need. No question. Thankfully, there are a number of hypotheses being looked at. I want to get Sara’s take. What do you think are some of the more promising strategies and some of the clinical trials that are ongoing to get us to the next level of benefit?

 A lot of the trials in the early disease setting are trying to say, “How can we improve outcomes for our highest-risk patients?” For example, the patients who have residual disease. Can we do better, potentially, then capecitabine is doing based on CREATE-X? We have trials that are comparing capecitabine with platinum therapy. That’s being done through the ECOG trial. I think we’ll learn more if there is a difference in outcomes.

There is a study that’s starting called the SASCIA trial, which is going to compare sacituzumab with treatment of choice in patients with residual disease. That’s particularly exciting because I think that sacituzumab seems to work really well in a refractory population. Maybe in this residual disease population that will have significant benefit. Then there are also studies that are looking at immunotherapy in patients with residual disease. This is going to be challenging because those trials are really focused on patients who didn’t get preoperative immunotherapy. We’ll learn if there is a benefit in the adjuvant setting in that population. Obviously this whole paradigm could shift if immunotherapy really does become part of the standard preoperative setting, as well as the adjuvant setting.

That was the way both IMpassion031 and KEYNOTE-522 still continued the I/O [immuno-oncology] even after surgery. The space is rapidly changing. There are also approaches about potentially trying to substitute different chemotherapies in the early disease setting. There’s even a study looking at sacituzumab in the preoperative setting. The idea is if we could see a really good pCR [pathologic complete response] with sacituzumab preoperatively, could this potentially even substitute for anthracycline as a backbone? We have a lot to learn, but definitely a lot of work is being done to see if we can improve outcomes compared with our current standards.

Thank you. We’ll have to stay tuned. We’ll have to have a lot more of these discussions over the year, so don’t go anywhere anybody. We’ll have to keep talking. There is a lot of promise, but the immunotherapy is definitely the thing we’re going to be watching most closely in the curative setting. Thank you, guys, very much for a great discussion.

Current Trials for Treatment of Early Stage TNBC

Sara, could you just mention this interesting study from ESMO [European Society for Medical Oncology Congress]? It’s more of an interesting point of biology and not really a practice-changing thing about TIL [tumor-infiltrating lymphocyte] as a biomarker in some of these earlier-stage TNBCs [triple-negative breast cancers].

 Absolutely. There are actually a fair amount of data continuing to mount data that for tumor infiltrating lymphocytes, the presence of them is associated with an increasing chance of pathologic CR [complete response] and a better outcome in terms of recurrence rate and maybe even overall survival, specifically for the triple-negative breast cancer subtype. At ESMO, there was this nice paper from the Netherlands Cancer Registry, in which they looked at patients over many years—a good number of patients with triple negative breast cancer who unbelievably did not receive any systemic chemotherapy.

They had tumor samples and looked back to see how these patients did based on the percentage of TILs in the tumor. They subdivided it for patients with 30% to 74% and then 75% and above, and showed that the patients with less than 75% TILs present, stromal TILs we’re talking about now, in the tumor had a much worse recurrence-free survival and overall survival. Those patients who had greater than 75% or 75% or greater were in the high 90% with no chemotherapy. The suggestion is maybe we could be using this as a marker to help select patients for a de-escalation strategy. These patients didn’t even receive chemotherapy, and their cancer is essentially cured because there is enough time of follow up to know that it didn’t come back. How to measure TILs is a point of contention. There is a lot of intralaboratory variability.

There now are international consensus scoring recommendations that are published that we can use, but it’s a bit like Ki-67 and the early days of measuring ER [estrogen receptor] and PR [progesterone receptor], and it’s kind of the Wild West. You have to be careful if you’re going to be looking at this and doing research with it. It’s not something we’re using prime time, but if it does become a way for us to select patients for a de-escalation strategy or for the use of immunotherapy, it’s going to have to be standardized.

Thank you. Right, very intriguing with a unique patient population that didn’t get any chemotherapy with node-negative disease and long follow-up. It’s really interesting, but as you said, maybe there could be a biomarker for de-escalation strategies for a future trial. Thanks, Sara.

Emerging Data on Biomarkers in ES TNBC

Let’s go to the other side of the coin. I’ve heard general enthusiasm for these data particularly in patients who are high risk and node positive. As Adam said, stage III; as Sara said, our highest-risk patients getting large deltas. That’s really important. It’s probably pretty compelling in practice, but let’s look at the toxicity. Adam, could you lead us off? What are some of the key findings on toxicity from the curative setting and the metastatic setting, for example?

I’ll do the curative, and then we can open the metastatic to everybody else. The bottom line is that in IMpassion031, just about everybody had an adverse event. The grade 3/4 was about 68%. One patient died on each arm of IMpassion031. Most of the adverse effects, at least the typical ones, were related mostly to chemotherapy. But when you start going into the adverse events of special interest, which are mostly the immune ones, the big 1 appeared to be—this is also in KEYNOTE-522, but I don’t have the data in front of me, but I remember it’s a little more—the hypothyroidism.

Any-grade hypothyroidism is 6.7% in the atezolizumab arm vs basically 1.2%. There was 3% hyperthyroidism, there was pneumonitis, and there was a little colitis but not a lot. Not as much as we’ve seen in the metastatic trials. Some diabetes and encephalitis. Rash seemed to be a pretty big one, but it was in both arms and infusion-related reactions. There was this ocular inflammatory toxicity, which I assume is a keratitis. The thing is that, at least to me, you saw most of these adverse events of special interest. My impression—and I’d like to hear my colleagues—is that these adverse events of special interest were there. They seem to be a little less frequent than in the metastatic setting, and the thyroid issues were what dominated this.

I remember this because at the Miami Breast Cancer Conference this year, I had a debate with Hope Rugo about this. We asked the audience, and it was really interesting what happened. Hypothyroid was a little higher in KEYNOTE-522. It was at least 10% or 15%. We asked is it OK to give a woman long-term hypothyroidism? I have women who got pembrolizumab in some of the trials, not on this 1 but on others, who are still getting thyroid hormone 3, 4, 5, 6 years later.

Is it OK to give someone long-term hypothyroidism? Especially someone who is young, who has triple-negative breast cancer. Yes, she’ll survive breast cancer, but on the other hand, now you’ve given her hypothyroidism. I just bring that to my colleagues. I don’t know the answer.

Thanks for that great summary of the things we’re facing every day with these immune-related adverse events. The question is, what’s your take on these in this setting? This is curative disease, long-term glandular toxicity, whether it is thyroid, whether it is adrenal; it is rarely pituitary. Some are probably reversible—hepatitis, colitis, pneumonitis are in the single digits but reversible. I want to hear you correct me if I’m wrong. The glandular stuff seems to be longer term, albeit rare, but these are serious.

Remember, we’re uncovering women. We forget we’re in endocrinology, right? We’re oncologists. We forget that. A lot of women float around with subclinical hypothyroidism, right? I mean, 15% to 20% of the population is the age that gets chemotherapy for breast cancer, immunotherapy for breast cancer. They have antithyroid antibodies. All we’re doing is giving them clinical hypothyroidism. So they have it anyway.

No doubt. That’s right. Exactly right where that might have come out anyway in a portion of patients. Kevin, what do you think? How do you balance this in the curative setting in these sometimes uncommon but pretty serious immune-related adverse events?

 It’s a real issue, and it’s why we need to be thoughtful in terms of who we’re considering giving checkpoint blockade. I’ll give you an anecdote because this is actively going on with 1 of my patients right now. This is a patient who was responding great to taxane-based therapy and was insistent on getting a checkpoint inhibitor. Somehow, with discussions with the family, discussing with a company, we were able to get the checkpoint inhibitor. She got 1 dose and developed the severe skin reaction and then dropped her hemoglobin down to 3 g/dL and developed a red cell aplasia. Twelve units of packed red blood cells later, she’s on high-dose steroids.

Just to put this in context a bit with what Sara Hurvitz was saying earlier. It takes just 1 patient who gets an anthracycline and has leukemia or a cardiac event for you to think, “Who should I really be using this for?” This is an example of how these drugs can have rare adverse effects, and when you see 1 it does give you a bit of pause. Even though it’s rare, you know, some of these toxicities can be long term. The other thing is that in I-SPY2, the rates of adrenal insufficiency were higher than what’s been reported as some of these other studies. The question is, is this just being monitored differently, captured differently, etc. I’m not underestimating the associated toxicities with these drugs.

Thank you. That’s very compelling. I want to hear from the Saras too. In terms of your gestalt, how do you balance this all? Let me start with Sara Tolaney this time.

: As Kevin was saying, just a couple of cases of significant immune-related toxicity always make you pause and really want to be very thoughtful about decisions about who gets treated. I’ve certainly had patients who have developed hypophysitis, thyroid abnormalities, and primary and secondary adrenal insufficiencies of both. It’s hard to see. They’re young women who are going to have this long term, and some of them are going to have lifelong steroids, and that’s a serious issue. In the metastatic setting, I’ve even had cases of type 1 diabetes develop. I have 1 patient now on an insulin pump. In a curative setting, this is much more of a lifetime impact, and we don’t know about fertility impact. Particularly in our young patients, we have no idea what that could be.

I do think we have to be very thoughtful about this, and that’s why in my mind, right now I’d want to restrict utilization to the high-risk patients, who need additional therapy that is providing significant benefit. In that case, it’s worth the potential toxicity risks that there are. As Kevin alluded to, these are rare things we’re mentioning. Obviously, if you look at the percentages, they’re not common. The challenge is we’re also still learning how to monitor for all this. We know to check the thyroid function all the time, but we’re not drawing routinely all the time. Cortisol is not in the guidelines. How many times may we be missing some adrenal insufficiency? We just need to be cautious.

Thank you. Sara, have the last word on this.

 I couldn’t agree more with my colleagues. The East and West Coasts are really coming together here. I completely agree. I shared an anecdote of a couple of patients I treated. One of them came in with these little lumps under her skin behind the knee, and I foolishly thought it was a Baker’s cyst—because I don’t remember general orthopedics or general internal medicine—and then it popped up behind her elbow. We had an ultrasound, and they were hard, hard masses that were growing. It was from 4 doses after pembrolizumab. It’s this granulomatous condition that can happen subcutaneously. This isn’t as serious as what Kevin and Sarah have shared, but clearly we are impacting their bodies and potentially in a harmful way. These patients are potentially cured with chemotherapy. I just want to underscore what everyone has already said. We have to take it very seriously, which is why we’re all very eager to see the EFS [event-free survival] and OS [overall survival] data come out.

Right. There’s really great consensus here. Let’s see some EFS, but for our highest-risk patients who we know we’re not making it, we’re not getting there with them. That’s a different matter because we know their prognosis is going to be abysmal. That’s where we’re going to be thinking about it. Thanks, everybody. Great conversation.

ICI-Based Therapy in TNBC: Immune-Related Adverse Events

 Kevin, take it away. Why are these studies different—NeoTRIP and IMpassion031—yet KEYNOTE-522 and IMpassion031 are very similar? Importantly, what do we do with these data now, Kevin?

 Many thoughts here. One, I would say in NeoTRIPaPDL1, the study was powered to look at event-free survival, so the study was not powered. One of the secondary objectives was to look at pCR [pathologic complete response]. I do think that with that study, a smaller study, and then KEYNOTE-522, which Sara had mentioned, we still need to follow the patients out. To me, just to summarize some of these things, in the KEYNOTE-522 and the IMpassion031, the things that were consistent between those 2 studies were that Adam and Sara had mentioned that PD-L1 expression can associate as being a predictor for response to chemotherapy but not necessarily predict who benefits from adding in checkpoint inhibitor. The other thing, I would say, is that when I was watching the presentation that was consistent between the 2 studies, those patients who had node-positive disease seemed to benefit the most with the addition of checkpoint inhibition.

We may not be seeing this benefit in NeoTRIPaPDL1 could be a few reasons. One of the things I proposed is there could there be a difference between PD-1 vs PD-L1 inhibitors, and maybe this had to do with blocking PD-L2. But clearly, we see a randomized study where there is a benefit with a PD-L1 inhibitor. So I don’t think that’s it. I also think that maybe this has to do with giving an anthracycline. I think 1 of the other studies that Sara Hurvitz had mentioned is the ubiquitous I-SPY2 study. They also had presented some data at San Antonio [Breast Cancer Symposium] in which patients received pembrolizumab plus a taxane followed by pembrolizumab alone without an anthracycline, and that was a negative arm. There may be some additional immune sensitivity with giving an anthracycline.

There is another thing I want to throw out. When I was watching the data being presented, in the arm in IMpassion031, which was a noncarboplatin-containing arm, we were seeing—not to compare studies—pCR rates that were relatively similar to what we saw in KEYNOTE-522, in which patients received a platinum. To me, it still asks this question: Do we give immunotherapy, whether patients really need to escalate with giving a platinum or not?

That’s another very, very important question coming out from these data. Go ahead, Adam.

I’m fascinated because, from a scientific point, why the node-positive ones are the ones that have benefited from the I/O [immuno-oncology] in comparison, and it’s telling us something. Maybe it’s telling us that those are the people whose immune systems aren’t dealing with the cancer, and that’s the 1 that needs the extra boost.

Of course, the cancer is sitting in the lymph nodes, so you’re starting to kill the cancer. It’s in the lymph nodes the TILs [tumor-infiltrating lymphocytes] are there.

This goes back to Bernie Fisher, right? The lymph nodes are just a marker that your whole body can or cannot deal with the cancer. It’s a radical idea, right? I don’t know if this is right or not, but we’re on this thing, so we can speculate and say whatever we want. The bottom line is that it maybe that’s what it is. Maybe it’s the fact that those are the people who need a checkpoint inhibitor because the fact that it got to the lymph nodes is telling you that your body may not be able to deal with it as well. I don’t know. I’m just raising that as a possibility.

Let me turn to the Saras and see what you think in terms of our practice. Of course, we wait for regulatory approval. We want the FDA to see all the data and look at everything that’s critically important. Nonetheless, we do have very, very high-risk patients in our practice. Even tomorrow, we have very high risk patients. Do we use them now? Do you think we’ll be using them? Also, what about this carboplatin question? It’s difficult to do cross-trial comparisons because the percentage of patients with nodes positive was not exactly the same. The KEYNOTE-522 trial, which had a higher-risk population, had more node-positive patients and, as I recall, also more stage III patients. It was a higher-risk population.

With the carboplatin, the pCR rate was 64.5% or something—not quite 65% but over 64%—vs just over 57%, so it was a little higher with the carboplatin in a more high-risk population. But close, as Kevin said, close. You’re not that far apart. Let’s get your perspective. We need to see and have more time to chew on it. Let me just ask Sara Hurvitz, what are your impressions?

The data—it seems like every day something else comes out, and that challenges the notions and the preconceived understanding of what’s happening. We make sense of it and then a study comes along and we have to rethink our paradigm thinking. We are thinking that perhaps paclitaxel isn’t the right thing to use with immune checkpoint inhibitors because you have to use steroids. We have several studies now that show increased pCR rates with paclitaxel. This anthracycline notion—there is something to it, Kevin. I hate to say it coming from the West Coast. I’m reluctant, I have to say, to use these routinely in all triple negative in the neoadjuvant setting.

However, in the neoadjuvant setting, if we’re seeing a patient not respond or frankly progress, this is where I’ve used it. I have to say, I’ve seen some very impressive responses to patients who I started on the taxane platinum, and the tumor did not budge. It’s clearly in the nodes, and we’re going in the wrong direction. I’ve switched to dose-dense AC [doxorubicin, cyclophosphamide]. I’ve been able to use the pembrolizumab in spite of it not being approved here, and I’ve had RCB-I [residual breast cancer stage 1] in 2 patients. I’m becoming a believer. I do think we really need to wait for the EFS [event-free survival] data before we use this routinely and need to have a better understanding of the biology and predictors of who really benefits.

Thank you. When you end up using it, just because you have tended to start with the taxane-carboplatin anyway, that’s probably going to be the paradigm you use hopefully, once we get full approval for these patients. How about you, Sara Tolaney?

 I’ve also been struggling. As Sara said, all these data keep coming out. I was en route with KEYNOTE-522 thinking, “OK, I know that there is an improved pCR, that this pCR difference is actually very large in the stage III patients.” The delta pCR is a little over 20%. It’s really a very big improvement. This is a population that has a very high risk of recurrence, and we do need to do something to improve outcomes. But where I’m a little weary is that there are—I’m sure we’ll get into more discussions about this—toxicities to weigh here.

Without a good biomarker to select which patients are the ones benefiting, we have to be very cautious in our patient selection. At this point we’re going to have to end up using anatomic risk because we don’t have a good biomarker. PD-L1 is not predicting who is benefiting. We see that the benefit is larger in these high-risk patients. Once we get an approval, that’s probably the direction I’d go: taking my node-positive patients as the patients that I probably would consider for adding immunotherapy. As everyone has alluded to, we really do want to see the long-term outcome data.

KEYNOTE-522 is powered for EFS. It was a very smart design to have a coprimary EFS end point, because that’s what we also really care about. Lack of pCR is associated with poor long-term outcomes. But on a trial-level basis, this is where we have not had good concordance definitively with PCR and EFS. We just want to see it. The trend is quite clear that it is heading there, but we do want it. IMpassion031 unfortunately was not powered for EFS. We’re never going to get from that trial a sufficiently powered EFS without the platinum backbone, which is a shame. They’ll get their data from their adjuvant trial.

That’s where it’s hard to make that decision, because we won’t be able to cross-trial compare EFS to know if we should be giving the platinum. It’s going to be based on risk, and I probably will avoid the platinum if I can because I feel like I’m already escalating therapy by adding immunotherapy. The delta was quite impressive in IMpassion031 without the platinum, with about a 16% delta. I probably would omit it. But in truth, it’s obviously not a data-driven decision because we don’t have that EFS to make that call.

Can I ask a question about a propensity match? This whole propensity matching was iffy 1 way or the other. But propensity matching across the 2 trials tried to find the patients who are most similar in the control arms of each to see whether the platinum is of benefit. I’m just thinking about an offhand way to figure this out.

Yeah, the numbers in IMpassion031 are going to be the things that are limiting. There will be data from the GeparDouze trial, which is preoperative. That does have carboplatin, paclitaxel, and AC in it, and that’s with atezolizumab.

We can talk about it later, but doesn’t B-59 require carboplatin also?

Yes, that’s what I was saying. There’s the B-59. There will be another big 1 that maybe will help. We’ll have enough numbers because we may have an apples-to-apples comparison there.

Role of Immune Checkpoint Inhibitor-Based Therapy in ES TNBC

 It was really weird. I’d love for Kevin to give a comment and maybe we’ll discuss this because he was the commenter on the NeoTRIP trial, and why it didn’t work. Because we have IMpassion031 now, and we knew it in a press release before ESMO. The idea behind IMpassion031 is they had 330 women with triple negative breast cancer, with a primary tumor greater than 2 cm. They had to have at least T2 disease; they could have N0 to N3 disease.

They had to have a known PD-L1 status, no prior treatment, and a decent performance status. They randomized them 1:1 to either atezolizumab with nab-paclitaxel weekly for 12 weeks, followed by atezolizumab with AC [doxorubicin, cyclophosphamide], Q2 [every 2] weeks. It was dose-dense [doxorubicin, cyclophosphamide]. So, it was weekly nab-paclitaxel, followed by dose-dense AC [doxorubicin, cyclophosphamide] times 4 with or without atezolizumab. It was placebo controlled. Then in the people who afterward, again there was a second randomization, but if you were randomized to the atezolizumab arm, you had 1200 mg intravenously Q3 [every 3] weeks for 11 doses. What was interesting is it was an adaptive design. They initially had started all comers, and then the IMpassion031 data came out, and they restricted it to the PD-L1 only patients. They required them to be PD-L1 positive in the trial, but the primary analysis was all 333 patients.

Looking at the results of the trial, it was interesting that overall in the study, the pathologic complete response rate was about 58%. It was 57.6% vs 41.1% overall, regardless of PD-L1 status. Joyce, you were saying what’s the baseline? There it is. That’s your baseline of all comers getting typical AC [doxorubicin, cyclophosphamide] or TAC [docetaxel, doxorubicin, cyclophosphamide] chemotherapy with the [docetaxel, doxorubicin, cyclophosphamide] being dose dense, is 41%. You did get, regardless of PD-L1 status, an improvement here of about 16% or 17 %. So afterwards, dividing it up by PD-L1 versus non PD-L1, the PD-L1 positivity was about 69%, it was 68.8% vs about 50%, 49.3% in terms of a pathologic complete response. Interestingly, and this is something that we really need to discuss, the patients who were PD-L1 negative, again SP142 less than 1%, assuming that meant they just had no TILs [tumor-infiltrating lymphocytes], is what I really believe it is.

Looking at that, it was 47.7% vs 34%. You’re getting a benefit here of about 13%, 14% in pathologic complete response from the checkpoint inhibitor in a PD-L1 negative subgroup of patients. I think that’s kind of important, looking at the hazard, plus it was across everybody, it didn’t matter. Again, inconsistent with the KEYNOTE-522 data, what they found is that in the atezolizumab arm, the event-free survival, when they checked, it was 10.3% had relapsed vs 13.1% with the placebo. This is in the median follow-up of about 20 months, and it was 4.2% in terms of overall death vs 5.4 %, a very small amount, but trending in the right direction. The hazard ratios were all about 0.7, 0.8, 0.6, something like that. The bottom line of this trial—and I guess we’ll talk about toxicity later, I won’t go into the toxicity, but just the results now—is that it looked like it was very similar in a way to KEYNOTE-522.

Now, why was it different from NeoTRIP, but the same roughly—we can argue about the details—as KEYNOTE-522? It’s really weird. I don’t want to steal your thunder, Joyce, but I’d love to hear from Kevin, again what he had said at San Antonio about why they were so different. Why would atezolizumab work in one of these and look kind of like KEYNOTE-522, kind of like pembrolizumab, but on the other scenario wasn’t? Is it patient selection? Is it the chemotherapies? What is it?

That’s great, super. Thanks so much, Adam. Those were really important data out of ESMO.

Early Stage TNBC: Phase 3 IMpassion031 Study

Let’s talk about immunotherapy because that’s really the next hope for our patients. Sara Hurvitz, maybe you could tell us about some of the data perhaps we already knew going into ESMO [the European Society for Medical Oncology Virtual Congress 2020], in terms of the immune checkpoint inhibitors in the early stage setting, and then we can hear from Adam on the new data that came out with IMpassion031 at ESMO this year.

: Absolutely. We’ve had a number of clinical trials that have reported out neoadjuvant treatment with immune checkpoint inhibitors. To just name a few, there was the KEYNOTE-522, the NeoTRIP trial, and the I-SPY study, which we hear about a lot of times because it has arms that open and close all the time. It just seems to be the ubiquitous I-SPY study. We also have the GeparNuevo study. Just to focus on the first 3, the I-SPY trial is an open-label, randomized phase 2 trial with adaptive design, and it’s evaluating stage II and III breast cancers of different subtypes.

Patients with HER2-negative disease, in this study, the particular one we’re talking about, were randomized to receive standard 4 cycles of weekly paclitaxel, followed by 4 cycles of AC [doxorubicin, cyclophosphamide] chemotherapy with or without pembrolizumab. Of course, pathologic complete response rate was the primary end point; they also looked at the RCB [residual cancer burden] index, and the 3-year event-free survival is being followed. In this study all arms or all cohorts graduated, meaning hormone-receptor positive, HER2 negative, and triple negative graduated to the next phase. They demonstrated at least an 85% predictive probability of success in a future phase 3 trial.

In this particular study, 29 patients with triple negative breast cancer were treated with the pembrolizumab compared to the control arm, and the pathologic complete response rate was higher, it was predicted to be higher, for the pembrolizumab arm. It was higher both in the PD-L1 positive as well as PD-L1 negative subsets. It’s interesting to note that higher PD-L1 expression actually is associated with a higher chance of achieving pathologic complete response, but regardless of the PD-L1 expression status, the use of pembrolizumab was associated with a higher rate of pathologic complete response. These data are very intriguing, but they’re certainly not large randomized clinical trial data.

Then comes along the KEYNOTE-522 phase 3 trial in triple negative breast cancer. This was a 2:1 randomization where patients received paclitaxel/carboplatin with pembrolizumab followed by AC [doxorubicin, cyclophosphamide]/pembrolizumab versus placebo with paclitaxel/carboplatin [doxorubicin, cyclophosphamide] chemotherapy. After surgery, all patients in the pembrolizumab arm continued to receive 9 cycles of pembrolizumab. Those in the placebo arm obviously don’t.

At the first interim analysis, and there are over 600 patients in this study, the pathologic complete response rate was about 65% associated with pembrolizumab, which was higher than with placebo. It did appear to be independent of PD-L1 expression. The event-free survival was trending toward improved at that early look, with very few events. It was in the right direction, which these very positive data underscore the reason it was published in 

. We were all getting really hyped up that it’s going to be here to stay, and we’re going to have an approval, and then comes along the NeoTRIP trial in which atezolizumab was looked at.

This was again a phase 3 trial, but much smaller, fewer than 300 patients, looking at carboplatin/nab-paclitaxel with no AC [doxorubicin, cyclophosphamide] given prior to surgery in this trial, it was all given on the back end. The use of atezolizumab here, although there was a numerical 2.5% or 3% improvement in pathologic complete response, there was no statistically significant difference. Even when you looked at it based on PD-L1 status, PD-L1 positivity did not predict for a better pathologic complete response by adding atezolizumab. So, the data are not totally consistent in the neoadjuvant setting, and there’s no FDA approval to date at the time of this recording.

Thank you. That was wonderful. Thanks so much for that excellent summary. We had the I-SPY and KEYNOTE-522 trials looking good, with nice deltas on pathologic complete response, and then unfortunately NeoTRIP was negative. There we were as we headed into ESMO.

Key Trials in ICI-Based Therapy for Early Stage TNBC

I want Sara Tolaney to have the last word on this, with regard to use of platinum and maybe even the node negative patient, the T1c N0. What do you do, Sara, if they don’t have a pathologic complete response [pCR]? How do we define pathologic complete response and what do we do if we don’t have one?

: I think in terms of the platinum issue, I feel very similarly to others that I generally tend to use ACT [doxorubicin, cyclophosphamide, paclitaxel] for the larger tumors. I only use platinum if someone is not having a good response generally to the upfront AC [doxorubicin, cyclophosphamide]. Then sometimes I’ll tack on the platinum because I have been concerned about the toxicity. We saw in the CALGB 40603 trial that the people who got platinum, if they had to not complete all their taxane, actually did worse. I do worry about adding on toxicity when we don’t have data that are sufficiently powered for long-term outcomes.

In the T1c tumors, I think like others have expressed, if they’re usually under 1.5 cm, sometimes I’ll do TC [docetaxel, cyclophosphamide] times 4. Whereas, if they start getting over 1.5 cm, closer to 2 cm, then I start giving them ACT [doxorubicin, cyclophosphamide, paclitaxel] therapy. I tend to use size to help me judge how much treatment to give them. I think generally speaking, we do tend to approach most of our patients with larger tumors with preoperative therapy, with the reason being that we can tailor their therapy based on response to that preoperative treatment. That tailoring of treatment is really going to change their long-term outcomes.

What we’re looking for is are they achieving a pathologic complete response to their preoperative treatment, meaning is there no invasive residual disease found within the breast or lymph nodes at the time of surgery? We know that those patients who achieve pCR have better long-term outcomes, whereas we know that those patients who have residual disease do have a higher risk of recurrence. As Kevin had alluded to, we do have data to suggest that we can alter those long-term outcomes for those patients with residual disease based on data from CREATE-X.

We know that in patients with residual disease who are randomized to get capecitabine vs no capecitabine, we saw that giving the capecitabine had a statistically significant improvement in disease-free and overall survival. We saw that that control arm with no capecitabine actually doesn’t do very well with the 5-year disease-free survival that’s around 55%. Again, there was a significant improvement to their outcomes by giving them capecitabine. I do offer all my patients with residual disease the 6 months of adjuvant capecitabine.

Sarah A. Hurvitz, MD, UCLA Jonsson Comprehensive Cancer Center

Articles