Andrew, let’s move now into probably the biggest news at the San Antonio Breast Cancer Symposium. I was involved in a panel, and we all had to go around and say what we thought was the biggest data at San Antonio—and this was unanimously selected, except by the main author, who is too humble to name it. Let’s talk about the RxPONDER trial. These data in lymph node-positive breast cancer have been long awaited, and the results were a little confusing. Again, this is 1 of the slide decks that I had to go through 2 or 3 times to really digest. Can you talk to me about this trial, its end points, how it differed from TAILORx, and what your conclusions were after seeing the data presented?

Yes, I’m not sure I found it as confusing as some, and I want to compliment Kevin Kalinsky MD, MS, a former Fellow from MSK [Memorial Sloan Kettering Cancer Center], who I couldn’t be more proud of for doing such a great job in communicating it. This was a much awaited study that probably will have the biggest practice impact change, where both pre and perimenopausal women with 1 to 3 positive nodes, about 5000 or so were randomized to endocrine therapy alone or endocrine plus chemotherapy. These women, importantly, they had to be considered to be able to get anthracycline and taxane therapy, so you are selecting out patients who passed that litmus test, and patients with recurrence scores from 0 to 25 were those randomized.

The overall result for the entire intent-to-treat population did not show the benefit for chemotherapy, but the preplanned subset analysis for premenopausal women did show a benefit in those patients. When you parse that further, based on subgroups of recurrence score 0 to 13, or 14 to 25, you see what you would expect, the signal for greater benefit the higher the recurrence score for chemotherapy. But even for premenopausal women in the recurrence score of 0 to 13 category, there was about a 4% benefit in invasive disease-free survival and about 6% for those with recurrence scores of 14 to 25.

The big high-level message here is that this is good news for a lot of postmenopausal women with N1 disease who now can safely forego chemotherapy. That’s a big sea change. One can always find subsets where you may have greater concern. The number of patients with, for example, 3 positive nodes was not huge. We don’t have information yet about the potential meaning of extracapsular extension, when you have those nodes that are significantly involved. Overall, again, from a very high level, this is an important trial that leads to meaningful de-escalation for a lot of patients.

Videos

Dr Andrew Seidman reacts to outcomes of the RxPONDER trial of chemotherapy use in women with HR-positive breast cancer at low risk of recurrence.

Results of RxPONDER in HR+ Breast Cancer

 Andrew, every patient is different, but how do you approach a low-risk, early-stage hormone receptor-positive breast cancer versus high-risk, nonmetastatic ER-positive breast cancer? Where are areas of unmet need?

 It all depends on how you define low risk, but if we take this at a very high level, patients with low-risk hormone receptor-positive HER2-negative cancer are those who will benefit from endocrine therapy, who don’t need chemotherapy. And those patients who are higher risk, either defined by anatomic features, as Sara pointed out, or molecular features, as Massimo pointed out, still need chemotherapy. I agree with Massimo’s point that chemotherapy is not the panacea and answer that we want it to be. And hopefully the next generation of trials for those patients with high-risk ER-positive disease will explore more rational therapeutics, those that are now being integrated in metastatic disease, for example, CDK4/6 inhibition, PI3 kinase inhibition perhaps.

Absolutely. I think it’s very hard for us to make a sweeping statement about how we treat each. For some patients we’re going to be using chemotherapy. We haven’t addressed whether ovarian suppression in a younger woman is enough and can supplant chemotherapy use, so many outstanding questions and many ongoing studies. Sara, a number of studies were presented and discussed at the San Antonio Breast Cancer Symposium. I’ve reviewed some of the slide sets I would say 4 or 5 times, and I’m still trying to grapple with the data and keep in mind the conclusions from each of these studies. Can you take us through some of the data that were presented at San Antonio relating to the use of biomarkers and assays to guide our treatment decisions for early stage disease?

 I absolutely agree, there were a lot of data that came out of San Antonio that were really interesting but also very confusing in light of different data sets, and really puts into question how we should be approaching our patients with ER-positive disease to make decisions about who actually needs chemotherapy versus not. One such study was ADAPT, which was a unique study because it was trying to look at if you assessed tumor response to endocrine therapy by Ki-67, could that really help you understand who needs chemotherapy versus not?

They did a unique thing, which was they integrated recurrence score and this response to preoperative endocrine therapy by Ki-67 to help address this question. They had 2 general groups of patients. They had patients who had a recurrence score that was low, so less than 11 and up to 3 positive nodes, who could get endocrine therapy.

But then they also took the patients with intermediate recurrence scores and gave them preoperative endocrine therapy, and looked at how they did with that to see what happened to their Ki-67. If they had a good response, meaning that they had a Ki-67 under 10% and went on to get endocrine therapy, how did they do compared to the people who had the low recurrence scores. Can we take an intermediate-risk patient and get away without chemotherapy if they respond to preoperative endocrine treatment? Is that a way that we can identify those patients who don’t need chemotherapy? In fact, they did see that invasive disease-free survival was noninferior between those patients who did achieve a low Ki-67 from preoperative endocrine treatment. So it was saying we could select patients who could get away without chemotherapy by looking at response to the short course endocrine exposure, which is an interesting way to think about things.

We also have another study that was trying to look at a couple questions, but one also that was trying to address looking at short-term exposure to endocrine therapy and seeing if we can just continue on endocrine treatment or escalate to chemotherapy. That was the ALTERNATE trial. This trial was asking a couple of questions. It was looking at if there are differences between different endocrine agents, specifically is there a difference between an AI [aromatase inhibitor] and fulvestrant, or fulvestrant with AI? It was trying to address, again, if there are differences between those agents in terms of tumor sensitivity to those endocrine treatments. But it was also looking to see, can we take those patients who get that short course endocrine treatment, and if they don’t achieve response to that endocrine treatment, should we escalate treatment to chemotherapy and then look at their outcomes?

We’ve seen now 2 presentations. We saw the first presentation at ASCO [the American Society of Clinical Oncology 2020 annual meeting], where we looked at the differences between the 3 endocrine agents, and there was really no difference between anastrozole and fulvestrant or the combination in terms of looking at PEPI [preoperative endocrine prognostic index] 0 scores. Many of us wondered if fulvestrant was better in the early-stage setting, and this suggests that there isn’t a difference between the 3 endocrine arms.

But then at San Antonio we saw what happened to those patients who didn’t get suppression of their Ki-67 from that window of exposure to endocrine therapy and went on to get neoadjuvant chemotherapy. What you see was that the pCR [pathologic complete response] rate for those patients was low. We know in general pCR to chemotherapy is unfortunately quite low, but it was only about 5% overall for that patient population. And most patients did have residual cancer burdens that were 2 to 3, so about 50% were RCB-2 [residual cancer burden-moderate] and about a quarter of patients were RCB-3 [residual cancer burden-extensive]. It was again suggesting that chemotherapy, as we know, doesn’t achieve high rates of pCR even in those patients who aren’t responding to preoperative endocrine treatment.

Then we saw lots of other data. One challenge we’ve had with using the genomic assays is that it provides you with some prognostic information, but it’s not individualized based on clinical pathologic features for the patient. If we send a recurrence score, it’s purely looking at gene expression in the tumor. It’s not factoring in how big that tumor was, to help you understand the absolute benefit that that patient could have from treatment.

So I think it was very interesting and helpful, the presentation that we saw regarding how to come up with a calculator, if you will, to help understand your individualized patient’s benefit. Joe Sparano, MD, presented this calculator that you could get online, which I did actually yesterday in clinic, to calculate for my patient. That was interesting because I think we’ve all wanted to merge this genomic information with our traditional clinical pathologic features, and it did seem like it was providing better information in terms of prognosis.

I think as you alluded to, we've got a lot of different pieces of information, which are all confusing when you put them together, about how we really should be making decisions for our patients. But I think it is teaching us that it’s more than just tumor size and nodal status to help us decide who needs chemotherapy versus endocrine treatment.

That was a great summary of some very complex data there. To add my 2 cents regarding the ADAPT study, I thought it was very elegantly designed. But the realism of testing Ki-67 in the neoadjuvant setting with endocrine therapy I don’t think is as simple as was presented, and I don’t think the uptake would be as good. We’re needing longer-term follow-up from these studies. I think we’re all begging for EFS [event-free survival] data, longer-term data to help us understand, for example, an alternate, whether chemotherapy is actually benefitting the patients who went on to do it, so doing randomized studies to query that.

I too have been using RSClin [recurrence score and clinical-pathological features]. I was involved in the “View From the Trenches” at the end of San Antonio, and Meredith Regan, ScD, put together this very nice table, different clinical scenarios and what the recurrence score would tell us about the patient, and the recurrence score adding in a couple features, and then using RSClin. And where RSClin I think is going to be most helpful is for those patients who have a score of like 26, 27 where we really don’t have good information about whether we should be using chemotherapy, and estimators for benefit from chemotherapy. I too have logged in and started using it a little bit in my patients to show them this pseudo individualized treatment response estimator.

Early-Stage HR+ Breast Cancer: Treatment Options

Let’s turn our focus now to the latter half of our program—an active and exciting, if not confusing, area—early stage hormone receptor-positive breast cancer. Let’s first take a high-level view.

Sara, can you talk to us about the rates of recurrence that we should be quoting to our patients with early stage hormone receptor-positive breast cancer who receive endocrine therapy? There are a lot of data, but a lot of it emanates from patients who were treated 10, 15 years ago, not necessarily with modern chemotherapy or modern regimens. What do we know about the risk of recurrence in a patient who comes in with early stage disease? What sort of features are associated with an increased risk of recurrence?

 Traditionally we’ve thought of very simple prognostic indicators based on clinical and pathologic features. First and foremost, we’ve always thought of tumor size, tumor grade, and nodal involvement as probably the most classic prognostic indicators that we have. But I think we’ve gotten a little more sophisticated as time has gone on, now understanding that we can potentially utilize molecular and genomic prognostic indicators to help better understand someone’s risk. We know, for example, that the 21-gene recurrence score could provide us with prognostic information in addition to predictive information. So the prognostic information that we get really helps us understand what someone’s risk of recurrence is. We also know Ki-67 is an assay that has been found to be prognostic, which, although a little more controversial, given some challenges with reproducibility of this assay.

From a very basic level, integrating clinical pathologic features and genomic/molecular prognostic indicators are probably our best understanding of how to put together someone’s potential risk of recurrence. I think the challenge in ER-positive disease is that biologically, these tumors can be very different. And so while we don’t classically utilize intrinsic subtype in the clinic, we do know that there are luminal B cancers that typically are higher grade, less endocrine-sensitive tumors, and then we sort of have the luminal B pathology, where there tends to be lower grade, more endocrine-sensitive tumors. How those cancers behave are just so different.

Not all ER-positive breast cancers behave the same, and so we see timing of recurrence is also different based on these different biologies, where we can have patients who have more aggressive luminal B cancers that are tending to be, again, these high-grade cancers that tend to recur earlier, and then tend to also see the reverse in patients who have more luminal B pathology where we can see very late recurrences.

I think we’re all trying to best understand how to figure out who’s going to recur and who’s going to recur early versus late, and there’s a lot of work being done in this field to tease this out further. But I think at a basic level, some of our prognostic factors really haven’t changed in terms of size, grade, and lymph node involvement. But now we’re getting a little trickier adding in these molecular and genomic features.

Excellent overview. Massimo, tell me which biomarkers and gene expression assays are available for our patients with early-stage disease. Can you sort out for us which are predictive versus which are prognostic markers?

Sure. The most important question we have been asking in the last few years is which patients we can de-escalate? When you think a patient is at high risk, you usually need to use your best therapy and your best chemotherapy possible. The Oncotype DX test has been used in node negative as well as node positive. Clearly, with a high Oncotype recurrence score, these patients need to receive chemotherapy. That may not even be enough, and I think this is the next level of care. For patients with an intermediate score, a lot depends on the clinical factors—the lymph node and the menopausal status. It seems like independently of what the recurrence score may be in this group, or independent of what the tumor size may be, there is a benefit from chemotherapy. This is something that probably needs to be discussed in more detail—what we should do with these patients when you have a molecular test that may not be so predictive.

Of course the alternative test, or the test has been used in the clinic, is the MammaPrint, 70 genes, which provides a score for high risk to low risk. It’s most of the time reported together with subtype, so you can have a luminal high risk, or a luminal low risk. Or you may find out if the disease is totally different, particularly in the group of patients who have very low expression. I think it is very useful to identify this as a luminal. But irrespective, I think as has been shown, when you have a patient who is genomic high risk, obviously there is benefit with chemotherapy. If you have clinically high risk in genomic 12, maybe this patient may be spared chemotherapy, and de-escalation is possible. There’s not a complete overlap between the two, but clearly most of the time they tend to indicate similar tumor characteristics.

Then the other question is the late recurrence. I think the EndoPredict test tries to define which patient needs standard endocrine therapy. I think this is less used and less known, but it’s a research question and a clinical question that probably we need to address now that we have all these molecular tests. I know there will be more and more effort to try to incorporate clinical issues, especially because women want to know. This is not a therapy that’s without adverse effects. They’re going to take 5 years of tamoxifen, and then 5 years an AI [aromatase inhibitor], with all the things that comes with it: hot flashes, weight gain, joint pain, and so forth. They’re going to take 10 years of tamoxifen, doubling essentially the exposure time to the drug. What if this patient develops a recurrence, how does this look molecularly in more aggressive and less aggressive subtypes? I think there will be a lot of use of this molecular test, again to de-escalate in this case the length of endocrine therapy when it’s possible, and maybe to add more combined therapy with the high risk.

Risk of Recurrence in Early-Stage HR+ Breast Cancer

Let’s round out this section now, Massimo, and talk a bit about resistance to CDK4/6 inhibitors. Can you tell us a little about what is known about the mechanisms of resistance to these? Give us some insights into what investigations are happening to attempt to address this clinical gap?

 The resistance to CDK4/6 has probably been studied with these phase 3 studies where we have collected plasma and done a number of retrospective analyses with regard to mutations. And also tissue analysis and expression was associated with resistance up front to Ibrance [palbociclib] in patients who have been exposed and in combination with fulvestrant. I think for a time line with regard to the genomic profiling with cell-free DNA, the persistence of PI3 kinase, 15 days and maybe at completion of treatment is a marker of resistance, one of the measurements of this particular hot spot mutation. Of course, we notice that there is a difference in the impact of the 2 mutations that are most common at the end of treatment. If they have onset of new mutation, you may think about resistance to fulvestrant and not necessarily resistance to CDK4/6, but nevertheless, we tend to.

We should start to differentiate genomic resistance into different buckets. At some point we may think about this as a situation of whether you change the endocrine backbone or continue? I think this is better addressed with clinical trials. We talked about the PACE study briefly. It was a randomized study looking at progression after first-line AI [aromatase inhibitor] and CDK4/6 with fulvestrant, combination with Ibrance, including a checkpoint inhibitor. There are a number of other studies looking at other pathways, for example, 

 mutation has been associated with resistance. There are agents that have been tested in combination at progression.

We talked about everolimus. That is probably 1 of the possible treatments in association with endocrine therapy because this pathway is very important with endocrine resistance overall, and maybe even CDK4/6 inhibitor resistance. There are a number of other cyclin inhibitors that have been tested. They’re going to be tested after CDK4/6.

Of course, the clinical question of using 1 CDK4/6 after the other, which happens in the clinic more frequently than not, based on the fact that maybe abemaciclib has single-agent activity. We can assume that after progression on chemotherapy or CDK4/6, there could be some efficacy of this agent, particularly in patients who may have developed brain metastases, we have the potential to prevent brain metastasis. Site of recurrence is important for the patient. For example, as we mentioned, we don’t have to forget the extent of disease—if the patient had extensive liver metastases, we do use another combination agent. We then go to chemotherapy. I think the community approach is probably using chemotherapy sooner than later. It is up to us to identify which patients would be resistant or which patient you can continue to use endocrine agents, single agent. I think the clinical trial is probably the answer. If somebody progressed on CDK4/6, and you do have an available clinical trial, please enroll the patient in a clinical trial with a targeted agent.

It’s clearly an area of unmet need and a lot of rich research that’s ongoing, so I would underscore your point about clinical trial enrollment in the second-line setting and beyond. I think it’s critical.

HR+ MBC: Resistance to CDK4/6 Inhibitors

I want to turn to a related topic but a bit of a different direction and talk about other therapies and approaches to targeting the PI3 kinase pathway that are emerging. Andrew, can you take us through some of the newer data, perhaps some data relating to capivasertib and other types of inhibitors like this, including ipatasertib, where we’ve had some negative data?

 For hormone receptor-positive, HER2-negative breast cancer, PIK3CA/AKT/mTOR is a frequently altered pathway and is often a key driver, so it’s natural to look at all components of this pathway. Of course, mTOR was the first that led to the clinical use of everolimus going back now almost a decade, and then PI3 kinase. The FAKTION trial, which was reported not that long ago, compared the use of fulvestrant plus placebo to fulvestrant plus capivasertib. And here there was a PFS [progression-free survival] advantage of about 5 to 10 months, favoring the use of capivasertib. This was a significant result, and the CAPItello trial, which is a phase 3 study, is ongoing to confirm that. Capivasertib is a potent selective oral inhibitor of all 3 isoforms of the serine/threonine kinase AKT.

 mutations are not very common, so we’ll have to see how big a piece of the breast cancer pie this ultimately impacts. Ipatasertib is another AKT inhibitor also being studied, and whether these agents will work best in combination with endocrine therapy or chemotherapy in triple-negative breast cancer is another unanswered question. There are many other drivers in that heterogeneous group of patients who are triple negative, and I think ongoing studies will have to inform us about that.

 I remember when everolimus first came out, there was a steep learning curve in managing toxicity with stomatitis, and now we’re using steroid mouthwash to help us mitigate these adverse effects. I feel really comfortable using everolimus. In the beginning I did not, but now I feel it’s like capecitabine, it’s fairly easy to manage. What have you experienced, Andrew, in the treatment of patients with alpelisib? What are the challenges you’re facing, and are you beginning to get to the top of that learning curve? How do you think clinicians are managing the unique toxicities associated with this particular inhibitor?

You’re absolutely right. The emerging toxicities from inhibiting this pathway have required us to be better internists. So going back to mTOR, having to relearn how to manage hyperglycemia and hyperlipidemia. For alpelisib, skin reactions are also important. The proactive use of antihistamines, of corticosteroids topically, is important. I work at an institution where we have an excellent support team in terms of dermatologists who are well versed in managing cutaneous toxicity of cytotoxic agents. I guess that would be the other new issue. When cutaneous toxicity does occur from alpelisib, it can be a deal-breaker for many patients.

Yes, absolutely. I think Hope Rugo, MD, presented some data regarding the use of upfront antihistamine. We’re now using that as routine as well. It’s funny, I had a very slender patient who I started on alpelisib, who had the lowest risk for hyperglycemia, and came in with a glucose of 550 mg/dL, very frightening. We have to watch these patients closely and make sure that we have our eye on the ball, and not send them off and say we’ll see you in 3 or 4 weeks.

I would say metformin is my go-to agent, the one I feel most comfortable with. My repertoire doesn’t extend too far beyond that, and if I need to, I start to engage my endocrinologists.

Targeting the PI3K/AKT/mTOR Pathway in HR+ MBC

The majority of patients won’t have a PI3 kinase pathway mutation. So let’s assume this particular patient has progressed on a frontline CDK4/6 inhibitor and has no mutation on analysis. Sara, can you take us through what sort of patient-related and tumor-related factors influence your treatment recommendation? What do we know about the use of other PI3 kinase pathway inhibitors in this situation?

If this patient fell into that 60% cohort without a PI3 kinase mutation who progressed on a CDK4/6 inhibitor, I think our choices here would be whether we feel comfortable using endocrine therapy alone, so for example, fulvestrant, or if we have a preference for adding a targeted agent to that fulvestrant, such as mTOR inhibition with everolimus? I think the third option, although it’s not one that I would utilize, but a question that arises would be, is there any benefit to continuation of CDK4/6 inhibition beyond progression? I think those are the 3 things that are often crossing people’s mind when they’re trying to figure out what to do.

How does one decide? Our data for adding mTOR inhibition to endocrine therapy predominantly come from the BOLERO-2 and PrECOG studies. BOLERO-2 looked at adding everolimus to exemestane, suggesting dramatic improvements in progression-free survival with a hazard ratio of around 0.4. Then we also have data from the PrECOG study showing very similar benefit for adding everolimus to a fulvestrant backbone. Obviously, the challenge here is these are data that came out prior to CDK4/6 inhibition being a standard first-line treatment option.

I think my general sense is that benefits from endocrine monotherapy, post-CDK4/6 progression, seem much more modest than they had been in patients who were CDK4/6 naїve. And given that, I have tended toward wanting to add additional targeted therapy to that endocrine backbone for my patients who may be wild type for PI3 kinase. So in this case my preference generally has been to add everolimus to fulvestrant in this type of a patient.

One could do fulvestrant alone as an alternative, and so when trying to decide that, I think the thing many of us are keeping in the back of our minds is, well, how well did that patient really do on endocrine therapy? Were they on their AI [aromatase inhibitor] CDK4/6 for 5 years and seemed very endocrine sensitive, or were they someone on it for 6 months where they progressed really quickly and I’m not feeling so comfortable about my endocrine treatment? I think these are factors that play into our decision-making about how comfortable we would feel with endocrine monotherapy in a patient, based on how well they’ve previously done with endocrine treatment.

In this particular case, again I think there are some who would give fulvestrant alone. I would probably give fulvestrant/everolimus. I typically have not utilized continuation of CDK4/6 beyond progression, and we do not yet have data for that from a randomized trial. There are trials ongoing, such as MAINTAIN and PACE, that will answer this question. But for now, what we have are case series, and I think they’re challenging to interpret because there’s a lot of selection that goes into a patient for whom you may continue beyond progression that’s going to bias your analyses. And these aren’t randomized studies. So we can’t really conclude from these series if CDK4/6 inhibition beyond progression is working. I think we need more work in that area, and again, hopefully in the near future we’ll get those answers.

I think you bring up really good points. The preclinical evidence that we have does not suggest that inhibition of CDK4/6 will benefit tumors beyond progression on one. I think some people are interested in seeing whether abemaciclib may have activity given it’s a more promiscuous inhibitor, perhaps hitting other targets and maybe would have differential activity. But I agree wholeheartedly with you that we need to see these clinical trials read out. I think we’ve gotten really good in the HER2-positive realm at switching out chemotherapy partners or HER2-targeted agents one after the other, but this is a situation where HER2 is the driver. So I think it’s important to distinguish that. That was a really nice review, and of course, we have the TRINITI study, which we’ve seen results from, but again not randomized, unfortunately, looking at ribociclib, everolimus, and exemestane altogether. So I think we do need the randomized data to come out.

Treating HR+ MBC After Frontline CDK4/6 Therapy

Thirty-five percent to 40% are going to have this actionable mutation 

 here. And if the liquid biopsy is negative, reflex to tumor tissue testing is important to pick up those false negatives, or where the mutation wasn’t selected.

Sara, what would you do for this patient? What treatment would you recommend for her, having progressed on an AI [aromatase inhibitor] and CDK4/6 inhibitor and having this mutation identified?

Given the SOLAR-1 data I would recommend fulvestrant with alpelisib. We know that it did have a significant improvement in progression-free survival [PFS] from 5.7 to about 11 months. We know that it improved response rates. And while there was not a significant improvement in overall survival, if you looked at the absolute improvement in patients with visceral involvement, such as this patient, it is quite striking—about a 14-month difference. So it seems like it is adding a lot of benefit. And while SOLAR-1 did not specifically look at patients with prior exposure to CDK4/6 inhibition, with only 20 patients in that trial, we now do have some data from the BYLieve trial to support that. Even though it was a single-arm study, compared to what we have historically seen for endocrine therapy alone, it does seem that adding PI3 kinase inhibition is likely adding a benefit in the CDK4/6 refractory population.

 Now Massimo, you had in our initial discussion talked about doing sequencing and genetic analysis of tumors at the first diagnosis of metastatic breast cancer, and some people are waiting for the time of progression. Let’s assume that you sequence this patient’s tumor right at the time she was diagnosed with metastatic breast cancer, and you know she has a 

mutation. Would this influence your selection of frontline therapy? Would you ever use alpelisib in the frontline setting, or would you still use a CDK4/ inhibitor in that particular patient?

That’s a good question, and I think regardless the PALOMA data looked specifically at the upfront mutations that were present in patients enrolled with endocrine resistance, that was a particularly resistant population. And we see the baseline 

 mutation in this population being common. It does not affect that because of the combination. We know that it’s absolutely appropriate to use the combination of a CDK4/6, in this case with an AI, otherwise fulvestrant, for patients in spite of the presence of a 

mutation up front. I think this is very important because it gives some sense of the type of sequence or treatment now that we have all these targeted agents. And as we talked about a patient who’s been exposed to CDK4/6, we certainly have the data that we could use a PI3 kinase inhibitor. We don’t have the data of the reverse, what if we used alpelisib first, and then we used a CKD4/6? I don’t think we can do the study because the data are overwhelming in regard to CDK4/6 with mutation. The data are available that show CD4/6 is effective for a 

Okay, great. Andrew, to round out this case now, let’s assume this woman had received instead fulvestrant plus a CDK4/6 inhibitor in the frontline setting. How would that influence the endocrine partner that you would use with alpelisib in this particular case? Do we have data to support the use of an AI with alpelisib?

Well, I guess I could say I believe that we do from the BYLieve trial data. So the endocrine partner can be an AI. I think the data that we have right now show better than expected results in terms of the proportion of patients who are free of progression at 6 months. As Hope Rugo, MD, had reported, roughly 50% median PFS of 7.3 months. There’s not yet randomized comparison to other approaches, for example, exemestane and everolimus. There was an interesting report, a real-world data study with Flatiron Health’s assistance, that compared outcomes of such patients to real-world choices, that did not involve alpelisib. The use of alpelisib with an endocrine partner following a CDK4/6 inhibitor yielded a PFS that was about double of whatever physicians were choosing to do in the real world. It’s nice to have options.

Treating PIK3CA Mutations in HR+ MBC

 Let’s turn to a patient who’s progressing after a CDK4/6 inhibitor. Unfortunately, patents do progress even though we’re seeing amazing PFS [progression-free survival] data. This is a 56-year-old woman who was diagnosed with a stage IIB hormone receptor–positive 

 [human epidermal growth factor receptor 2]–negative breast cancer. She had breast-conserving surgery, received adjuvant AC [doxorubicin, cyclophosphamide]–paclitaxel followed by radiation therapy, and an adjuvant aromatase inhibitor [AI] for 5 years. Three years after completing the adjuvant AI, she developed abdominal and hip pain. She had staging scans that showed metastatic recurrence in the liver and bone, which of course is confirmed by biopsy.

She received an AI plus a CDK4/6 inhibitor as frontline therapy and 29 months later developed progression in the liver and bone. She has an ECOG performance status of 0. No lab abnormalities were noted. She did have the ctDNA [circulating tumor DNA] sent for PI3 kinase analysis, and she has a 

Andrew, what’s your approach to molecular testing in this situation at the first progression? We’ve, of course sequenced her here. Are you just looking for 

 mutations? Are you doing the whole next-generation sequencing on a liquid biopsy or tumor?

 Certainly many people use the theracreen liquid biopsy assay for 

-mutation status. At MSK [Memorial Sloan Kettering Cancer Center], we have an in-house platform called MSK Access to do this. We can get other information, so I often do have that information. Not that I’m using it for off-clinical-trial action ability. But this is the moment absolutely for someone who has progressed on AI and CDK4/6 inhibition to be thinking about determining whether PI3 kinase is mutant or wild type. I often do start with liquid, and if it’s wild type and there’s easily accessible tissue, I may also get a biopsy at that time.

 mutations, how common they are, and hormone receptor–positive breast cancers? What should our process be if the liquid biopsy does not indicate a 

 mutation? Should we be reflexing to tumor tissue?

The case that you indicated certainly has been exposed for quite some time. There’s a liver that can definitely be biopsied, and I would certainly go for that: Confirm the yes status, and probably send for NGS [next-generation sequencing].

I will probably use an NGS liquid biopsy to have other information. Not necessarily only a 

 mutation. This is a driver mutation. It’s present in 30%, 35% of patients with metastatic breast cancer. It’s 1 of the most common occurrences. Because a driver mutation means it’s been there, and it’s probably not always driving the resistance to endocrine therapy.

The activation of the pathway—all the way down to mTOR—has been associated with endocrine resistance. The importance of that is PIK3 inhibitors, which proved to be manageable in terms of toxicity. We probably will talk more about this regarding toxicity.

In the SOLO-1 trial there was advantage over the androgen therapy, where we have a mutation and tissue also in the liquid biopsy.

I would think it is an important mutation for actionable treatments in patients with resistance. That is very common in metastatic breast cancer. Not always. We need to act on mutations present.

Case Discussion 1: Resistance to a CDK4/6 Inhibitor

Andrew, we have MONALEESA-3 and MONALEESA-7, which have shown overall survival benefit with ribociclib in the frontline setting. We’re still waiting for PALOMA-2 and MONARCH 3 in the frontline setting to read out in terms of survival. Does this impact your selection and the use of a CDK4/6 inhibitor in the frontline setting, or do you think these are all fairly similar drugs? Which of them are appropriate?

 Some of us are old enough to remember the wars of the aromatase inhibitors. Is anastrozole better the letrozole? Is exemestane better than either? You know, Coke vs Pepsi, paclitaxel vs docetaxel wars. Happily, we’re not spending our time and energy doing these trials, comparing CDK4/6 inhibitors. As Massimo pointed out, the hazard ratios for PFS [progression-free survival] benefit are remarkably constant; population differences make it problematic to do cross-trial comparisons.

My own view, shared by a working group that I chaired for the NCI [National Cancer Institute] Breast Cancer Steering Committee, is that PFS still has primacy in the first-line setting for ER [estrogen receptor] positive 

 [human epidermal growth factor receptor 2] negative. Our working group reported our consensus findings in the 

Oncology] a couple of years ago in a special article.

There are certainly many chapters to be written afterward. While it’s really nice to have OS [overall survival], and it’s compelling, whether OS is going to be driven by the event rate or the statistical power, you get the key value. My long-winded answer to your short question is that these 3 agents are much more similar than they are different in terms of efficacy. OS is nice to have but not needed in the first-line setting.

Excellent. I’d like to ask you to expand a little on the toxicity profile of these 3 agents. Which toxicities do you find most challenging? Which do you think are consequential and inconsequential and really not much of a bother in how this impacts your selection?

That’s an important question because if one accepts that efficacy is more similar than different, then the only reason to pick and choose would be perceived differences and likely toxicities. So the commonality across the 3 CDK4/6 inhibitors in clinical use is the potential for neutropenia. But we feel pretty comfortable knowing how to manage that now by checking CBC [complete blood count] twice weekly in the first 2 months, perhaps at the first day of each of the next couple of cycles, then based on discretion afterward. The same applies for potential liver function abnormalities.

The issues of abemaciclib related to GI [gastrointestinal] toxicity distinguish it a bit from palbociclib and ribociclib. With experience and time and the use of antimotility agent’s physicians have really become comfortable using this agent, notably in the MONARCH trial the likelihood of discontinuing abemaciclib for diarrhea, for example, was quite low. And with ribociclib there’s still the need to do ECG [electrocardiogram] monitoring every 2 weeks in the first month to look for QT prolongation.

 The points you brought up are really the key toxicities we’re seeing. We have to overlay the fact that we are in the midst of a horrendous pandemic. Is that insulin staying in your selection of agents in terms of ones that have less neutropenia associated? Is that playing any role in how you select?

Yes, with grade 3 and 4 neutropenia—cross-trial comparisons were lower with ABEMA [abemaciclib] than with RIBO [ribociclib] and PALBO [palbociclib]. I’m not sure I consciously choose CDK4/6 inhibitor differentially because of the pandemic or make across-the-board dose reductions or tell patients to stay home and not come in for necessary monitoring.

We can all be afraid of COVID-19 [coronavirus disease 2019], and we should be. But our patients fear the cancer that they have, the known rather than the unknown.

Absolutely. I agree. Massimo, when you’re treating a patient with the CDK4/6 inhibitor, are there particular patients for whom you’re most worried about toxicity? Who are patients that, as Sara described, really are the highest risk where you might consider not using a CDK4/6 inhibitor because of their situation, their characteristics?

As we said, comorbid conditions may drive the choice. We haven’t seen an increase in infections, including viral infections with CDK4/6. In clinical practice I have seen in our clinic some cases of COVID-19 irrespective of the CDK4/6. So I would say the neutropenia probably doesn’t increase risk of COVID-19; it’s an independent prior infection. Even the length of infection seems to be relatively independent disease. If you’re a younger patient, maybe you have symptomatic disease. Just to close the chapter on COVID-19, that should not be a driver for change.

But we have seen patients who have comorbid conditions but are unable to get to the hospital because of COVID-19. Patients with inflammatory bowel disease were begging me not to give a drug that would produce diarrhea. I didn’t do it. They responded, of course, to Ibrance without having a problem.

There are some other patients who were started on palbociclib and developed a neutropenia very quickly. We had to drop the dose to 75 mg.

CDK4/6 Inhibitors for HR+ MBC: Data and Utilization

All the features you bring up are important when we’re evaluating a patient. In addition to looking at comorbidities, they help predict how they’re going to do overall. But as we begin to go into the clinical trial data, it’s probably important for us to distinguish what’s prognostic and what’s predictive of benefit from a therapy.

Massimo, take us through the phase 3 data relating to CDK4/6 inhibitors in the frontline setting, supporting the use of these agents in the frontline setting. We have 3 inhibitors. Take us through the 4 clinical trials that have frontline patients. I may be miscounting. There seems to be a lot relating to the use of these drugs and the differences in these clinical trials, highlighting the overall survival that we’ve seen in the frontline setting and what we’re waiting to see.

 The CDK4/6 inhibitors have completely changed our view on how we manage patients with hormone receptor–positive disease. For the first time, we started to see an overall survival benefit. We’ve never thought that this was going to be possible.

There are a number of studies—MONARCH 2, MONALEESA-7, MONALEESA-3—looking at the combinations of endocrine agents. Most of the time, Faslodex or a combination of hormone blockage with standard endocrine therapy has shown that compared with standard endocrine therapy, there is an improvement in survival even though they have expression in the front line. We’re starting to look at the differences in the studies to see the different populations.

Also, this was the first Alliance study to use an AI [aromatase inhibitor] or fulvestrant in the patients in the metastatic arm. In the cases where this was perimenopausal, premenopausal, there was a group of patients also in the first line. They were not exposed, but the majority were mostly endocrine resistant.

There are some studies where the patients were endocrine resistant by definition. Some others included de novo, in which the patient has not been exposed to prior endocrine therapy. These characteristics are important when we are using it in clinical practice. Other factors have been brought up and analyzed when we look at the specific studies: site of metastasis and their setting of endocrine resistance in the disease. It seems like some of the agents would be more effective in endocrine resistance or visceral metastases. Some of the others are particularly effective in endocrine sensitivity and bone disease. So it’s very difficult to figure out if this is an agent difference or if it is a characteristic of the patients enrolled in the study. The 3 agents may be similar.

When you look at the data and, particularly when you are in front of a patient, you need to decide. All this will eventually play a role. For example, if you have a first line, maybe de novo and bone disease only, there’s no question that using palbociclib in combination with an AI is absolutely a reasonable choice. If you have premenopausal patients who develop metastatic recurrence or have not been treated, ribociclib pretty much is supported by the data.

At the same time, if you have a primary-resistant disease, maybe with visceral metastases or liver disease, abemaciclib is something that may make them more comfortable. This is not necessarily because of the difference in the agents but primarily because of the data focus or reach of this particular population. Obviously toxicity profile plays a role. How many times have we seen an older woman and maybe their bone marrow is not perfect? You don’t want to de-escalate Ibrance or overdose. So you will just say 1 agent instead of the other because some other patients have more sensitive GI [gastrointestinal] symptoms, and you feel comfortable using Ibrance instead of abemaciclib. It’s great that we have all these agents and these options, because all of them have great efficacy in combination advantages.

Excellent. Sara, are there any patients for whom you would not use a CDK4/6 inhibitor in the frontline setting? Have you run into that in the last year or so, where you say it’s just not appropriate or you don’t think they’ll benefit?

There are very few patients for whom that’s actually happened. As Massimo was saying, there’s the benefit of PFS [progression-free survival], and now we’re seeing signals for survival. The benefits are so compelling that you really do want to give patients CKD4/6 inhibition. The challenge though is that 1, we can’t find a patient who’s not benefiting from therapy when we look at all the subgroup analyses. In general, everyone is deriving benefit. But it’s a little more challenging to take therapy with the CKD4/6 inhibitors than it would be to take endocrine treatment alone. We obviously know there are patients who do well with endocrine treatment alone, sometimes on an AI for years, but we can’t figure out who those patients are to know up front whom we can omit the treatment from.

I want to try to give it to patients, but they have to come in more frequently for lab checks. There are some older patients I have who just can’t get into clinics or a lab to get their blood drawn. The comorbidities are such that it just would be challenging. So it’s a rare patient. We know the older patients benefit, just like the younger ones, but they do have a little more in terms of toxicity. It is something that we have to factor in—just the practicality with their other comorbidities and how challenging it is for them to come in for more frequent evaluations.

Yeah, the forest plots are pretty striking across the different clinical trials, so you can’t pick out a subgroup who doesn’t appear to benefit. The comorbidities are a key point, as is the toxicity profile that Massimo brought up as well.

Frontline Therapy for HR+ MBC: CDK4/6 Inhibitors

®, “Expert Perspectives on Advanced Hormone Receptor–Positive Breast Cancer.”

I am Dr Sara Hurvitz, from the David Geffen School of Medicine at UCLA in Los Angeles, California. Joining me today in this virtual discussion are my colleagues Dr Massimo Cristofanilli from Northwestern University Feinberg School of Medicine in Chicago, Illinois; Dr Andy Seidman from Memorial Sloan Kettering Cancer Center in New York, New York; and Dr Sara Tolaney from the Dana-Farber Cancer Institute at Harvard University in Boston, Massachusetts.

Today we are going to highlight a number of topics pertaining to systemic treatment for HR+ [hormone receptor–positive] breast cancers and the impact of recent clinical trial data on clinical decision-making, including data from the 2020 San Antonio Breast Cancer Symposium. Let’s get started on our first topic.

What we’re going to do right now is focus primarily on CDK4/6 inhibitors in hormone receptor–positive metastatic breast cancer. To lead us in this discussion, I’d like to ask Massimo: What is your approach to molecular testing for patients with hormone receptor–positive metastatic breast cancer at the time of initial diagnosis? Are you doing molecular testing in the frontline setting?

Thank you, Sara. This is an important question now that we have so many targeted therapies and so much information about the recent genomic data with regard to specific therapies, specifically resistance to therapies.

I would differentiate the 2 drugs that we deal with mostly. One was a de novo disease, because de novo and metastatic disease are usually what we have seen in 30% to 40% of the patients. In this case, we really don’t need genomic therapy because we already know the standard of care. There is no mechanism of resistance that we’re going to target up front, and we’ll talk more about the role of CDK4/6 information in these patients.

The patient has a recurrence, and that’s interesting because they may have a recurrence while they’re receiving adjuvant therapy. This primary resistant setting, and it’s very important to find out the genomic drivers of resistance. Obviously when we treat a biopsy in the metastatic setting, if there is tissue available from additional sites, we send for genomic testing and NGS [next-generation sequencing] to find out if there are, for example, 

If there is bone disease, we can repeat NGS [next-generation sequencing], so we probably look at liquid biopsies and wait to understand the molecular driver for these patients. That would be very important to decide, for example, the endocrine therapy backbone, which can be more effective in these patients.

So you are doing molecular testing in a patient whose disease has recurred on an aromatase inhibitor. Do you want to see whether there’s 

 mutation in order to determine which endocrine partner you’re going to use in the frontline setting? Is that the reason you would do that?

Correct. In some cases we have seen, even more for cancer, we may even find 

 mutations that are actionable at this point. There are data suggesting this patient responds to tyrosine kinase inhibitors. But it’s very important to rule out immediately if the 

 mutation is the driver of the resistance. As you know, there are agents being developed in this space, especially after Faslodex or maybe in parallel as randomized studies, to find the best agent.

Endocrine resistance is becoming more of a complex issue that needs to be investigated; not 1 single mutation can be responsible. What we have is a mutation and the 

Andrew, do you agree? Do you do molecular testing in the frontline setting, or are you waiting until the patients have progressed on frontline therapy to evaluate their tumor biology?

 Besides considering patients for clinical trials in the first-line setting of novel agents like PI3 kinase inhibitors, all I really need to know to treat patients are ER [estrogen receptor], PR [progesterone receptor], and 

 [human epidermal growth factor receptor 2] status. Generally, and certainly for the de novo patient, I don’t do additional molecular testing at that time. There are reasons to think about germline 

 testing for subsequent use of PARP inhibitors, but even in the first-line setting that’s not top on the menu.

Most patients are getting it done at some point. There may be some differences in the timing. Some people wait until second- or third-line setting. Certainly for clinical trials, many of us are interested in querying what the tumor biology is for clinical trial eligibility.

Sara, let’s turn to prognostic factors that you look at when you’re evaluating a patient in the frontline setting with hormone receptor–positive metastatic breast cancer. How do you weigh different prognostic features? What features do you look at when you’re determining a therapy?

In early stage disease we have discreet prognostic indictors—whereas, in metastatic disease, it has been generally a little more diffuse. When I’m trying to assess prognosis in someone with either recurrent disease in the first-line setting or de novo disease, 1 important feature is disease-free interval. If someone has recurred from prior adjuvant therapy, how long ago was that adjuvant treatment? Did they recur on their endocrine therapy? Did they recur off it? How long off their therapy had there been recurrence? These are really important features because we know that people who are recurrent on their endocrine therapy—particularly early on in their endocrine treatment, usually within the first 2 years of adjuvant therapy—are patients with endocrine-refractory disease. Those are patients who we know are not going to be as sensitive to future endocrine therapy and generally do have poor prognosis. Someone who recurs 10, 15 years later or who has de novo disease—those are patients with very different biology in their tumor and generally have better prognostic features.

We also try to look at sites of disease, so we know that patients who have significant amounts of visceral involvement generally have poorer prognosis than someone with, for example, bone-only metastases. So lots of liver metastases are certainly a worse prognostic feature. There are also some data to suggest that PR negativity can be a poor prognosis, and patients who have higher-grade tumors as well. Those are things we generally have kept in mind when trying to think about prognostic features.

There are also other molecular tests that could be prognostic. Massimo has done a lot of work in this area looking at CTCs [circulating tumor cells] that has been shown to be prognostic. It is an area that is a little more diffused than in the early sage setting, where we’re trying to factor in lots of pieces of the puzzle to best understand how those patients are going to do.

Andrew D. Seidman, MD, Memorial Sloan Kettering Cancer Center

Articles