New Treatment Standards Spark Questions About De-escalation and Sequencing in Breast Cancer

Sara M. Tolaney, MD, MPH

Several news standards have been introduced throughout breast cancer, including the addition of pembrolizumab (Keytruda) to chemotherapy in early-stage triple-negative breast cancer (TNBC), the incorporation of abemaciclib (Verzenio) into the adjuvant setting in hormone receptor–positive breast cancer, and the use of olaparib (Lynparza) in BRCA-mutated, HER2-negative, high-risk disease, explained Sara M. Tolaney, MD, MPH.

She added, however, that although each approach has been practice changing, there are many lingering questions with each new standard, such as whether platinum is necessary in early-stage TNBC, whether Ki-67 should guide use of abemaciclib, and whether abemaciclib or olaparib should be given in the presence of a germline BRCA mutation.

“As drugs emerge, it does create a lot of questions about how to optimally place them in our current treatment paradigm, because the data are changing so fast. There isn’t always a perfect answer for what the right step is,” Tolaney said in an interview with OncLive^® following an Institutional Perspectives in Cancer webinar on breast cancer, which she co-chaired.

In the interview, Tolaney, chief, Division of Breast Oncology, Susan F. Smith Center for Women’s Cancers, associate director, Susan F. Smith Center for Women’s Cancers, senior physician, Dana-Farber Cancer Institute, associate professor of medicine, Harvard Medical School, reviewed practice-changing data sets throughout breast cancer and provided perspective on potential treatment approaches in the absence of definitive results.

OncLive^®: Your presentation focused on treatment in TNBC. What strategies have shown benefit in patients who are at high risk for recurrence?

Tolaney: In general, patients with TNBC, unfortunately, are at higher risk of recurrence than patients with other disease subtypes. We have tried to augment outcomes for these patients by looking at different add-on strategies. Some of these strategies have been increasing the amount of chemotherapy we’re giving, so in addition to anthracycline/taxane therapy, trying to add platinum, which has been shown to increase pathologic complete response [pCR] rates. There’s some controversial suggestion that [adding platinum] may also improve event-free survival [EFS].

We have also seen data looking at adding immunotherapy to chemotherapy that have consistently shown improvements in pCR rates by adding immunotherapy to anthracycline-based chemotherapy. There are now several other novel strategies that are being considered as add-on strategies, such as even looking at antibody-drug conjugates [ADCs] in this early disease setting. A lot of improvements are being made for our patients with early-stage TNBC, which is great.

Could you elaborate on the work that has been done to incorporate platinum into chemotherapy backbones? Is such an approach necessary for all patients?

Several randomized trials have looked at adding platinum chemotherapy to an anthracycline and taxane backbone in the preoperative setting and have shown improvements in pCR rates. For example, we saw this in data from the GeparSixto [NCT01426880], CALGB 40603 [NCT00861705], and BrighTNess [NCT02032277] trials. However, these trials have reported different EFS data. GeparSixto showed that adding platinum improved EFS. CALGB 40603 suggested no improvement in EFS, and BrighTNess showed an improvement in EFS. It’s challenging because these were preoperative trials that were powered for pCR not EFS. It’s a little [difficult] to understand whether platinum improves EFS.

In truth, while not really knowing the answer to that question, [adding platinum] has really become our new standard chemotherapy backbone given the data from KEYNOTE-522 [NCT03036488]. We still need to tease out how much chemotherapy we really need to be giving patients and whether what we’re doing is too much. There will be other suggestions to help us figure this out in the future. An NRG trial that was conducted had randomized patients in the adjuvant setting to an anthracycline/taxane, with or without carboplatin. We’ll have to wait for those data because that study, being adequately powered, will help definitively answer whether platinum improves survival outcomes.

Based on the data from the phase 3 KEYNOTE-522 study, has pembrolizumab (Keytruda) plus chemotherapy become the standard of care for all patients with early-stage TNBC, regardless of response to neoadjuvant therapy?

KEYNOTE-522 looked at adding pembrolizumab to anthracycline, taxane, and platinum chemotherapy, and we saw very clearly that adding pembrolizumab improved pCR rates with the final delta being around 7.5 between the 2 arms when you looked at all approximately 1100 patients in the trial. It’s interesting because after surgery, all the patients who got pembrolizumab preoperatively continued the pembrolizumab in the adjuvant setting, irrespective of what happened at the time of surgery in terms of response, which certainly isn’t the way we’ve been practicing where we adapt adjuvant therapy based on response to preoperative treatment. There are several controversies about how to handle the adjuvant setting, but nonetheless, what these data show is that adding pembrolizumab in this fashion improves pCR rates and significantly improves EFS.

Given these data, preoperative pembrolizumab with chemotherapy is the new standard of care for all patients with stage II and III breast cancer because we found that the benefit was seen across all subgroups. Even in the lower-risk groups of the overall population––the node-negative patients and the stage II patients relative to stage III patients––the hazard ratio for benefit was similar for pembrolizumab.

Biomarkers haven’t proven useful to date with PD-L1 not being a biomarker of benefit to pembrolizumab. Currently, I have been standardly offering pembrolizumab with anthracycline, taxane, and platinum-based chemotherapy in the preoperative setting to these patients with stage II and III disease.

You also mentioned ADCs in your presentation. How might we see the use of these agents evolve in TNBC?

I have been really excited about ADCs because they offer a very clever way to deliver a lot of chemotherapy in a targeted manner to cancer cells. We’ve seen that this approach is successful in the metastatic setting with sacituzumab govitecan-hziy [Trodelvy] paving the way in metastatic triple-negative disease, demonstrating an improvement in progression-free survival [PFS] and overall survival [OS] in the ASCENT study [NCT02574455].

What is interesting is ASCENT enrolled a population of patients who were chemotherapy-refractory. These were patients who had already progressed through multiple lines of chemotherapy before they got sacituzumab govitecan. [The results showed that] sacituzumab govitecan was better than getting standard chemotherapy in that setting. Could you then suppose that if you took a patient who got preoperative chemotherapy and had residual disease, demonstrating they’re chemotherapy-refractory: Could you then use sacituzumab govitecan in that setting, and would that be better than our standard capecitabine approach, which is currently used? That is an ideal setting to think about ADCs in; it’s also a setting we do need to do better in.

Patients who have residual disease are at high risk of recurrence. We’ve currently been using capecitabine, and, now in the era of KEYNOTE-522, many of us will use capecitabine with pembrolizumab for our BRCA wild-type patients. Can we do better than that? An ADC may be a new way to [improve the standard].

Agents like sacituzumab govitecan, maybe even datopotamab deruxtecan, another TROP2-directed ADC, could be something to consider in this refractory setting. I also wonder about agents like fam-trastuzumab deruxtecan-nxki [Enhertu], which we’re now seeing benefit with in HER2-low breast cancer. For the one-third of patients, for example, with triple-negative disease who are HER2 low: Could trastuzumab deruxtecan have a role? We’ll see a lot more from these ADCs hopefully soon.

What would you like to stress regarding the management of TNBC?

Unfortunately, TNBC has a bad name. It’s really defined by what it isn’t: lack of the estrogen receptor, progesterone receptor, and HER2, but we’re learning that there are receptors on TNBC that we can target that can allow us to personalize therapy for these patients. For example, we know if you have PD-L1 positivity, immunotherapy can work. We now have ADCs that target different receptors on TNBC. Will we need to know if someone is HER2 low, for example, to use a HER2-directed ADC? Will TROP2 expression ever have a role in helping us select patients for ADCs in the future?

We used to think of TNBC as a cancer that didn’t have a target. We were just using standard chemotherapy. Now we’ve seen a whole new armamentarium of therapies, not just immunotherapy but ADCs and PARP inhibitors. This field is dramatically changing, which is nice to see.

Adrienne G. Waks, MD, of Dana-Farber Cancer Institute, spoke about updates in HER2-positive breast cancer, highlighting the phase 3 DESTINY-Breast03 (NCT03529110) and phase 2 HER2CLIMB (NCT02614794) trials. How have these studies affected how treatment sequencing is thought about in the metastatic setting?

It’s an exciting time in metastatic HER2-positive disease where there are so many good choices. It makes it hard to make [treatment] sequencing decisions, which is a very good problem to have. I was very excited to see the data from DESTINY-Breast03, showing that trastuzumab deruxtecan is dramatically better than ado-trastuzumab emtansine [T-DM1; Kadcyla] as a second-line and beyond approach in the metastatic setting. [DESTINY-Breast03 resulted in] probably the best hazard ratio I’ve ever seen from a phase 3 trial of around 0.28, favoring trastuzumab deruxtecan. The study established unprecedented outcomes for this ADC in metastatic HER2-positive disease and has established it as a second-line standard of care.

Now there’s work trying to move trastuzumab deruxtecan earlier with DESTINY-Breast09 [NCT04784715], which is looking at trastuzumab deruxtecan in the first-line setting and DESTINY-Breast05 [NCT04622319] in the post preoperative setting. A lot more to come from this ADC in metastatic disease.

On the flip side, we also have data with tucatinib [Tukysa] from HER2CLIMB, another practice-changing study, suggesting not just a PFS benefit but also an OS benefit even in patients with active brain metastases. We have never seen a registration study with OS benefit in patients with active brain metastases. It does bring up a bit of a conundrum of what to do, for example, if you have a patient who may have both agents as options. What we’re putting together is that given the great results from DESTINY-Breast03, trastuzumab deruxtecan makes the most sense in someone without active brain metastases, but if someone has active brain metastases, thinking about tucatinib and then sequencing trastuzumab deruxtecan later makes sense.

Dr Waks also touched on treatment for patients with early-stage HER2-positive disease and how it has become more individualized. To what do you attribute that?

We’re coming a long way in HER2-positive early-stage disease where we are figuring out how to tailor therapy to the individual patient. We’ve figured out that we can tailor [treatment] based on response to preoperative therapy. If someone has a pCR, then we can continue monoclonal antibodies, but if they have residual disease, given the data from KATHERINE [NCT01772472], escalating [therapy] by using T-DM1 instead of monoclonal antibody therapy is dramatically better, with a hazard ratio of about 0.5, suggesting a 50% reduction in invasive disease-free survival [iDFS] events with T-DM1.

On the flip side, we also have options to de-escalate therapy. We’ve tried to focus on de-escalation by selecting patients who have less high-risk clinical pathologic features and by tailoring [treatment] based on response to therapy. If the patient achieves a pCR: Can we get away with less? Could you use, for example, a chemotherapy-light backbone in a preoperative setting, and if the patient achieves a pCR, say that chemotherapy was enough for that patient? Even though it [would be] less than our standard approach, could we say it’s enough and continue the monoclonal antibody?

There are several trials now trying to address these de-escalation strategies. One such study, CompassHER2-pCR [NCT04266249] is looking at only giving a taxane with trastuzumab [Herceptin] and pertuzumab [Perjeta; HP] for 12 weeks. If you get a pCR, you continue HP maintenance in the adjuvant setting. That is a great step in the right direction, allowing us to de-escalate treatment based on pCR.

Also having agents that are friendlier to patients makes a big difference in de-escalation, such as subcutaneous HP, also known as Phesgo, which has been great to have in the clinics because the injection allows patients to get in and out of infusion so much faster rather than waiting through sequential, monoclonal antibody treatments. Now ongoing studies are looking at home administration of Phesgo, which is exciting to think that maybe in the future, we’ll be able to figure out ways to allow patients to self-administer the drug, which would really be a tremendous improvement in patients’ quality of life.

In the presentation Philip D. Poorvu, MD, of Dana-Farber Cancer Institute, gave on hormone receptor–positive, HER2-negative breast cancer, he spoke about the monarchE trial (NCT03155997) with abemaciclib. Has abemaciclib become a standard approach for patients with high risk?

monarchE looked at adjuvant abemaciclib in patients with high-risk, hormone receptor–positive, HER2-negative disease. In the study, high risk was defined by 4 or more positive nodes or between 1 and 3 positive nodes and a tumor that was either greater than 5 cm, high grade, or had high Ki-67. That study showed that giving 2 years of abemaciclib in combination with endocrine therapy resulted in a 30% reduction in IDFS events, so a significant improvement in outcomes for these patients with hormone receptor–positive, high-risk disease. It is standard now to offer patients who meet these high-risk criteria adjuvant abemaciclib.

It is interesting though, that although the intention-to-treat [ITT] population for monarchE was this group of high-risk patients, the FDA decided to limit the approval for abemaciclib to patients who had those high-risk [inclusion] features, which was a bit of surprising, because we did see in monarchE that the benefit for abemaciclib was seen in both the high Ki-67 as well as low Ki-67 patients. Ki-67 was not a predictive biomarker for benefit with abemaciclib but rather a prognostic biomarker. Other guidelines have come out both from the American Society of Clinical Oncology and the National Comprehensive Cancer Network, suggesting that you can use abemaciclib in the ITT population. That’s how our group [at Dana-Farber Cancer Institute] has been practicing, using the all-comer population for selecting patients for use of abemaciclib.

What might explain the differences in outcomes between monarchE and PALLAS (NCT02513394) and PENELOPE-B (NCT01864746)?

In the metastatic setting, we’ve seen remarkably similar PFS data emerge from all 3 CDK4/6 inhibitors, whether it’s palbociclib [Ibrance], abemaciclib, or ribociclib [Kisqali]. It was a bit surprising to see differing results for these agents in the adjuvant setting. We saw data from PALLAS and PENELOPE-B, which had explored the use of adjuvant palbociclib and had not shown an improvement in IDFS from the addition of palbociclib to endocrine therapy. That was quite shocking to me. PALLAS had looked at 2 years of adjuvant palbociclib in patients with stage II and III disease about 60% of whom would have met monarchE eligibility. Despite this, we didn’t see benefit even in the high-risk subgroup.

There was a lot of confusion about why palbociclib has benefit in patients with metastatic disease, just like the other agents, but doesn’t seem to have benefit in the adjuvant population. Was it something funny about the trial design, rather than it being the agent itself? We can’t really tease that out because it wasn’t just that the investigators selected low-risk patients. Even when we look at the 60% of patients who were high risk, we still don’t see benefit. Some people argued that maybe patients weren’t getting enough drug because there were very strict criteria for discontinuation if you had neutropenia in that trial. About 40% of patients didn’t complete the 2 years of palbociclib, but even when the investigators looked at dose intensity, they didn’t see that dose exposure influenced outcomes.

We’re still a bit puzzled about that and wonder whether there are differences in the CDK4/6 inhibitors themselves. Is there something different about these agents that leads to differences in efficacy in an early-stage population? Palbociclib is given intermittently, whereas abemaciclib is given continuously. Does taking the brakes off the cell cycle with that week off affect efficacy? Is it that [palbociclib] is a less potent CDK4/6 inhibitor than abemaciclib? We don’t know. PENELOPE-B validated the fact that palbociclib doesn’t have benefit [in the adjuvant setting] because it too didn’t show benefit.

We are now awaiting data from NATALEE [NCT03701334], which looked at 3 years of adjuvant ribociclib in an intermediate- and high-risk patient population. That study has completed accrual and will be interesting because it may address some of these questions, such as whether it’s the CDK4/6 inhibitor itself [that’s causing disparate outcomes] and whether there are differences [in the agents] if ribociclib shows benefit where palbociclib hasn’t. Could it also be the duration where [NATALEE is evaluating] 3 years of therapy whereas PALLAS and monarchE evaluated 2 years of therapy. There are a lot of unknowns here about duration and differences in these agents but, at least for now, we do have abemaciclib as an option for patients with high-risk, hormone receptor–positive breast cancer.

Where do you see selective ER degraders (SERDs) and ADCs playing a role in this space?

We’re also seeing several new agents for patients with metastatic hormone receptor–positive breast cancer, and we hope that many of these agents can also move into the early-stage setting. Some examples include things like oral SERDs, where we now have data from the EMERALD study [NCT03778931] in the metastatic setting suggesting that elacestrant was better than endocrine therapy of choice in a pretreated, metastatic hormone receptor–positive breast cancer population. It does make one wonder whether these oral agents are better than our adjuvant endocrine therapy choices like aromatase inhibitors [AIs] and tamoxifen. There are now several adjuvant trials [evaluating oral SERDs] that are enrolling. One such example is an adjuvant study looking at giredestrant, an oral SERD that’s being looked at in an immediate adjuvant population. There’s also amcenestrant, a SERD, which is looking at amcenestrant in an AI-intolerant population in a bit of a later setting. A lot more to come with these agents.

There are also ADCs. A recent press release came out that findings from TROPiCS-02 [NCT03901339] showed sacituzumab govitecan was better than chemotherapy of choice in hormone receptor–positive metastatic disease. Another press release came out stating that trastuzumab deruxtecan is better than chemotherapy in patients with HER2-low metastatic breast cancer, the majority of which in this trial were hormone receptor positive. It’s very promising to see these press releases. It begs the question: Can these ADCs have benefit in earlier-stage patients, and should we be exploring that as well?

What questions remain in the field?

A lot of questions come up about how to approach patients who have triple-negative, early-stage breast cancer and residual disease. This is a little bit controversial right now, because in KEYNOTE-522, the standard was everyone got adjuvant pembrolizumab, whereas prior to KEYNOTE-522 reporting, our standard approach was if you had residual disease and were BRCA wild-type, we gave you 6 months of adjuvant capecitabine based on data from CREATE-X. Now we also have data from OlympiA [NCT02032823], suggesting that in patients with germline BRCA mutations, 1 year of olaparib is beneficial.

What do you then do when you have these different data sets that emerged separately over time? What do you do in the current era? The truth is, we don’t have optimal data to address this question. Patients with residual disease are unfortunately still at high risk, even after immunotherapy-based treatment, so many of us are giving capecitabine with pembrolizumab to the BRCA wild-type patients. If patients have germline BRCA mutations, we’re giving them adjuvant olaparib in combination with pembrolizumab.

Another question that comes up in patients with high-risk, hormone receptor–positive disease and a germline BRCA mutation is: What are you going to give? Are you going to give olaparib, or are you going to give abemaciclib? That’s a tough question too, where, again, we don’t have any data to tell us what the right answer is. If you look at the monarchE data, we saw a 30% reduction in IDFS events, but we don’t have this subgroup data to know what happened to the patients with germline BRCA mutations in that population. Are they receiving similar benefit as the ITT population with abemaciclib? How does that compare with the 40% IDFS reduction and OS benefit seen with olaparib?

No one knows the correct answer here, but my general approach has been to use olaparib in the patients with germline BRCA mutations, because there’s both DFS and OS benefit, and the benefit is robust and biologically makes a lot of sense. Could one even sequence abemaciclib, potentially post olaparib? Theoretically, yes. Is there data for that? No. But could you think about it for someone who has multiple positive nodes who you’re really worried about is very high risk? Yes, one could consider that with careful monitoring of adverse effects.