What about approved options for relapsed or recurrent HR-positive/HER2-negative MBC in the second- and later-line settings? This is a situation where we are increasingly moving into treatment personalization. This slide shows a general treatment algorithm following disease progression on first-line ET. 

We also need to know whether patients have a PIK3CA mutation. If so, they are candidates for alpelisib plus fulvestrant provided their hemoglobin A1C is <6.5%.^5,27

In addition, FDA approval of capivasertib for patients who progress on a CDK4/6 inhibitor is anticipated in the near future.²⁸ It is not yet known whether capivasertib approval will be targeted to patients with an alteration in the PI3K/AKT/mTOR pathway, which would include PIK3CA and AKT mutations, as well as PTEN loss, or whether capivasertib will be approved for an all-comer population.

As part of the decision-making process, we need to know whether patients have an ESR1 mutation. This is an acquired mutation that develops after progression on an AI or AI plus CDK4/6 inhibitor in the metastatic setting. ESR1 mutation is not typically found in primary breast cancer, and it is often present in the first metastatic biopsy in only approximately 5% of patients.²⁹ By contrast, up to 40% of patients will develop an ESR1 mutation after progression on first- or second-line ET.^29,30 For individuals with an ESR1 mutation, elacestrant, an oral selective estrogen receptor degrader (SERD), is a potential treatment option.⁵

We also need to know whether disease harbors somatic or germline BRCA or PALB2 mutations. In select cases, one could use olaparib or talazoparib based on available data.^5,31 

We also would want to know if the relapsed MBC has developed an acquired HER2 mutation. If so, there are favorable data from the SUMMIT trial showing substantial activity with the triplet combination of neratinib, trastuzumab, and fulvestrant in patients with HER2-mutated—not amplified—HR-positive/HER2-negative MBC.³²

In addition, we want to know whether the tumors have a high tumor mutation burden (TMB-H) signature or can be classified as microsatellite instability–high (MSI-H)/mismatch repair deficient status. These biomarkers are uncommon in HR-positive/HER2-negative MBC, but if they are present, we can potentially offer the patient treatment with single-agent immunotherapy (either dostarlimab or pembrolizumab).^5,22,33

In later lines (≥4), once patients become ET resistant, we often introduce capecitabine and, thereafter, move on to one of the ADCs, including T-DXd if the disease is HER2 low (defined as IHC 1+ or 2+ but FISH negative) or sacituzumab govitecan if the disease is HER2 IHC 0.⁵

Support for use of the oral PI3K inhibitor alpelisib in patients with PIK3CA mutations comes from the randomized phase III SOLAR-1 study.^34,35 Patients in this trial were randomized to receive alpelisib plus fulvestrant or placebo plus fulvestrant. It is important to note that the SOLAR-1 trial participants had not previously received a CDK4/6 inhibitor, as that was not yet a standard of care at the time this trial was conducted. The data showed a significant improvement in median PFS with the addition of alpelisib to fulvestrant (11.0 vs 5.7 months; P = .00065).³⁴ There was also a trend toward improved OS, although this was not statistically significant (median OS: 39.3 vs 31.4 months; P = .15).³⁵

PIK3CA mutations, which occur in approximately 35% to 40% of cases, can be detected in the primary breast cancer, in metastatic tissue, and in ctDNA to identify individuals who may be candidates for alpelisib.³⁶ However, suitable patients need to have good glycemic control, with a hemoglobin A1C <6.5%, given the potential for hyperglycemia associated with alpelisib.^27,34

BYLieve was a large phase II trial that enrolled patients whose HR-positive/HER2-negative, PIK3CA-mutated breast cancer was progressing through first- or second-line treatment.^37,38 Cohort A included patients progressing on an AI plus CDK4/6 inhibitor, and these individuals received treatment with alpelisib plus fulvestrant. After a median follow-up of 21.8 months, the median PFS and OS were 8.0 and 27.3 months, respectively. On the bottom right, we can see the primary analysis response data showing that the CBR was 45% and the ORR was 17% in patients with confirmed PI3KCA mutation. Of note, median PFS is not as good as that reported in SOLAR-1 (11.0 months). However, post CDK4/6 inhibitor, a median PFS of 8.0 months with fulvestrant and alpelisib is clinically meaningful for patients.

It is also important to mention the ongoing phase III CAPItello-291 clinical trial, which evaluated the investigational oral AKT inhibitor capivasertib in the second-line setting in combination with ET. As stated previously, capivasertib is not yet approved by the FDA, although we do anticipate FDA approval in the second half of 2023 based on the promising results of this study.

CAPItello-291 was conducted in men and postmenopausal women with HR-positive/HER2-negative MBC who had progressed on first-line AI therapy, and they could have previously received a CDK4/6 inhibitor (N = 708).²⁸ In fact, ≥51% of patients were required to have a CDK4/6 inhibitor before trial entry. Eligible individuals were randomized to receive either fulvestrant with placebo or fulvestrant with capivasertib 400 mg orally twice daily administered at a schedule of 4 days on and 3 days off. The 3 days off allows patients to recover from toxicity, particularly diarrhea, and this strategy appears to work well. 

There were dual primary endpoints: PFS in the overall population unselected for any mutations in the PI3K/AKT pathway and PFS in the cohort of approximately 40% of patients who had a PIK3CA mutation, AKT mutation, or PTEN loss-of-function alteration. 

The primary PFS results are shown here. The median PFS for the overall population on the left was 7.2 months with capivasertib vs 3.6 months without, and the median PFS for the AKT pathway–altered population on the right was 7.3 months with capivasertib vs 3.1 months without.²⁸ The hazard ratios were 0.60 and 0.50, respectively, in the overall and AKT pathway–altered populations. These were both statistically significantly (P <.001), so the data are very encouraging. It also is important to note that the PFS improvement with the addition of capivasertib to fulvestrant was consistent across all subgroups assessed, including those patients who had previously received a CDK4/6 inhibitor. 

At the time of this analysis, the OS data are immature, but the survival curves look encouraging. Survival is trending toward improvement with capivasertib, with hazard ratios of 0.74 in the overall population and 0.69 in the AKT-altered population.²⁸

 I am hoping for the FDA approval of capivasertib in the overall population so that we can offer this agent to a broader group of patients whose disease is progressing while they are receiving a CDK4/6 inhibitor.

What about the safety profile of capivasertib? Approximately 13% of patients discontinued capivasertib because of AEs, which is not that high compared with other targeted agents.²⁸ The predominant AE was diarrhea; 72.4% of patients receiving capivasertib experienced any-grade diarrhea. Most diarrhea was grade 1 in severity, but a small proportion (approximately 9.3% of patients) had grade 3 diarrhea. Diarrhea tends to show tachyphylaxis (acute and short-term onset of drug tolerance) over time. Taking those 3 days off capivasertib each week means the diarrhea is usually short-lived.

Some nausea was associated with capivasertib (all grades: 34.6%; grade 3: 0.8%). Rash also occurred (all grades: 22.0%; grade 3: 5.4%). However, within the study, advisors did not prophylactically use nondrowsy antihistamines, as we do with alpelisib and everolimus, so that will be interesting to look at in the future. Of importance, 16.3% of patients had hyperglycemia, which differentiates capivasertib from alpelisib, where hyperglycemia is an on-target effect.

In summary, the main toxicities with capivasertib are diarrhea, nausea, rash, and fatigue.²⁸ However, those 3 days off during each week of dosing make a big difference in terms of allowing patients to stay on therapy. Dose reductions occurred in approximately 20% of patients, so that is always an option for patients to continue on therapy. 

I also am hoping for capivasertib to become available by the second half of 2023. Right now, I do use alpelisib and fulvestrant in patients whose breast cancer has a PIK3CA mutation.

In patients with recurrent disease and ESR1 mutation, we have elacestrant as a treatment option. As a reminder, ESR1 mutations develop as a resistance mechanism to ET, particularly AIs, in the metastatic setting.²⁹ Approximately 40% of patients will develop an ESR1 mutation over time under the selective pressure of ET.^29,30 Because of this, we must assess ctDNA upon progression on first-line ET and also upon progression on second-line therapy.⁵ ctDNA is the best way to find ESR1 mutations. 

In the randomized, open-label phase III EMERALD trial, eligible patients had received treatment with a first-line CDK4/6 inhibitor and ET (N = 478).³⁹ Upon progression, they were randomized to receive either elacestrant at 400 mg once daily or investigator’s choice of single-agent fulvestrant, anastrozole, letrozole, and exemestane. The coprimary endpoints were PFS in the all-comers population and PFS in the ESR1-mutated population, which comprised approximately 48% of the population. 

In the EMERALD trial, both the overall population and the population with ESR1-mutated disease yielded a statistically significant improvement in PFS for patients treated with elacestrant, meeting the primary endpoint.³⁹

Shown here are data that can help inform how we in the clinic can best choose which patients should receive elacestrant. The table on the bottom shows the PFS outcomes in the population with ESR1-mutated disease based on previous PFS with a CDK4/6 inhibitor of ≥6, ≥12, and ≥18 months, which helps illustrate the likelihood of ET sensitivity given that elacestrant is administered as a single agent and pinpoints where the agent has most activity (if ≥12 months), resulting in FDA approval. Patients who were receiving a prior CDK4/6 inhibitor for ≥12 months (in the middle of the bottom table) or ≥18 months (on the right in the bottom table) had a median PFS with single-agent elacestrant of 8.6 months compared with approximately 2 months with single-agent ET.³⁹ So, that is the sweet spot—patients with ESR1-mutated disease who were receiving a previous CDK4/6 inhibitor for ≥1 year—as this increases the likelihood that they still have some degree of ET-sensitive disease.

If we take a look at the table on the bottom for patients with ESR1-mutated disease, those individuals who had a shorter duration of previous treatment with a CDK4/6 inhibitor did not benefit from treatment with elacestrant. Patients who were receiving their previous CDK4/6 inhibitor for <6 months had a median PFS of 1.9 months with elacestrant, which was the same as with standard therapy.³⁹ Similar PFS durations were seen for both elacestrant and standard therapy for individuals who had been receiving a previous CDK4/6 inhibitor for 6-12 months. Thus, these data provide a good rationale for offering elacestrant to patients who attained ≥1 year of benefit from their previous CDK4/6 inhibitor—which is the vast majority of patients, as the median PFS with first-line CDK4/6 inhibitor use is 2-3 years.⁴⁰ 

The safety profile of elacestrant is excellent. Patients can expect low-grade nausea (grade 3 in 2.5%), which can be prevented by taking the medication with food.⁴¹ If patients have dyspepsia, they can take a proton pump inhibitor because elacestrant does not need an acidic environment to be absorbed. Approximately 35% of patients in the EMERALD trial had any nausea; however, patients in the trial were not taking the medication with food or a proton pump inhibitor.^30,39 Only 8% of patients on the elacestrant arm required use of an antiemetic, and only 3.4% of patients discontinued elacestrant because of AEs, which reaffirms that it was a very well-tolerated therapy.

Elacestrant is an important agent for us as healthcare professionals. I choose it for patients who have disease with a confirmed ESR1 mutation that I think is ET sensitive and who were receiving their previous CDK4/6 inhibitor for ≥12 months. For those patients with more indolent disease, I have a window of opportunity where I have 3-4 months to see if the patients’ disease has a nice response to elacestrant. I do not use single-agent elacestrant in patients with disease that is rapidly progressing in the liver or with lymphangitic spread.

Other promising oral SERDs are currently in development. One such agent is the investigational SERD camizestrant. Here we show the study design for the phase II SERENA-2 trial evaluating 3 doses of camizestrant in patients who experienced disease progression with first-line AI with or without a CDK4/6 inhibitor (N = 240). Patients were randomized to receive camizestrant 75 mg, 150 mg, or 300 mg orally daily vs fulvestrant 500 mg IM.⁴² The 300 mg was stopped early with a clinical study protocol amendment, and efficacy is being evaluated only with the 75-mg and 150-mg doses. The primary endpoint of this trial was investigator-assessed PFS. Secondary endpoints included CBR at 24 months, ORR, OS, and safety.

The data reported by Oliveira and colleagues at the 2022 San Antonio Breast Cancer Symposium showed PFS improvements for both the 75-mg and 150-mg doses of camizestrant vs fulvestrant (median PFS: 7.2-7.7 vs 3.7 months; P <.02 for both doses).⁴² This PFS benefit was consistent in patients who had received a prior CDK4/6 inhibitor, who comprised more than 50% of the population. Moreover, the secondary endpoints of ORR and CBR at 24 months also were improved.

Thus, camizestrant is clearly superior compared with fulvestrant, although they are both SERDs. Data analyses are ongoing to look at efficacy in patients with ESR1-mutated disease vs the all-comers population, so we will stay tuned for those results.

Data from the SERENA-2 study show that camizestrant was very well tolerated.⁴² In general, patients report few noticeable AEs. A minority of patients experience an AE called photopsia, which is basically a transient change in light sensitivity. When patients go from a very dark room into bright sunlight, for example, they can feel that the light is particularly bright, like a flash of light. However, it does not get in the way of daily life and, fortunately, is reversible when patients stop camizestrant.

Otherwise, this is a very well-tolerated therapy, and we look forward to definitive phase III trial results. There are several ongoing phase III trials with the potential for establishing camizestrant in the HR-positive/HER2-negative metastatic space. At this time, camizestrant use remains investigational.

We also are keeping an eye out for another very promising investigational oral SERD called imlunestrant. I want to briefly discuss data from the phase Ia/b EMBER study, where imlunestrant was combined with abemaciclib. Patients with ER-positive/HER2-negative ABC were randomized to receive imlunestrant (either 400 mg or 800 mg orally once daily) plus abemaciclib with or without an AI (N = 85).⁴³ Patients on this trial were stratified based on menopausal status and the presence of visceral metastases. Patients in this trial had not previously received a CDK4/6 inhibitor, but they did have up to 1 prior line of ET. The primary endpoint of this study is determining the recommended phase II dose.

Pooled PFS, on the right, was very favorable with the imlunestrant/abemaciclib combination, and the addition of an AI did not appear to make much difference in terms of further improving PFS.⁴³ In both arms, 80% of patients were progression free at 1 year.

The CBR was also very good at approximately 75% among all patients on the trial and was roughly similar in the 2 arms.

Like other oral SERDs, imlunestrant was very well tolerated.⁴³ Inclusion of abemaciclib does, of course, bring diarrhea into the picture. We are well aware of that, and we can dose reduce abemaciclib to get the diarrhea down to a grade 1 or grade 0 level. Grade 1 diarrhea usually can be managed with loperamide. Really, however, the toxicities seen here are from abemaciclib and not from the imlunestrant. 

Overall, these data are very encouraging, and we look forward to the results of the global phase III EMBER-3 trial, which is evaluating imlunestrant alone vs imlunestrant plus abemaciclib vs investigator’s choice of ET in patients progressing on first-line AI with or without a CDK4/6 inhibitor (NCT04975308).

Unfortunately, there comes a time when the disease becomes resistant to ET, and we have to go on to cytotoxic therapy. Most of us will use capecitabine first in the HR-positive/HER2-negative population, but then we have to go on to IV cytotoxic therapy, with 2 main options available, including ADCs. Whereas T-DXd is preferred for the HER2-low population, sacituzumab govitecan is preferred for the HER2-zero population.⁵ 

The phase III TROPiCS-02 trial led to FDA approval of sacituzumab govitecan, an ADC with SN-38, an irinotecan-derivative payload that is targeted toward an ubiquitously expressed tumor-associated antigen called TROP-2.⁴⁴ The patients on TROPiCS-02 were very heavily pretreated, having received 2-4 lines of prior chemotherapy. Participants were randomized to receive sacituzumab govitecan 10 mg/kg on Day 1 and Day 8 of a 21-day cycle or to receive single-agent chemotherapy. The primary endpoint was PFS. 

Results from the TROPICS-02 clinical trial show that median PFS was significantly improved with sacituzumab govitecan vs chemotherapy (5.5 vs 4.0 months; P = .0001). Of importance, we also saw a significant improvement in median OS (14.5 vs 11.2 months; P = .0133).⁴⁵ There was a 3.3-month improvement in median OS with sacituzumab govitecan vs chemotherapy, and we see a greater percentage of patients alive at 12 months (60.9% vs 47.1%) and 18 months (39.2% vs 31.7%).

Survival trumps everything, so these data show that sacituzumab govitecan is more effective to give to our previously treated patients with HR-positive/HER2-negative ABC compared with single-agent chemotherapy.

When analyzing the results according to TROP-2 expression level, we see that expression of this marker does not make a difference in terms of efficacy.⁴⁵ Patients benefit regardless of the degree of TROP-2 expression, so—thankfully—we do not have to select patients based on TROP-2 expression level.

The most common any-grade AEs of sacituzumab govitecan are neutropenia (71%) and diarrhea (62%).^45,46 Some patients also may experience alopecia, and scalp cooling does help at least somewhat with the alopecia. 

The prescribing information recommends premedication to prevent likelihood of infusion reactions and chemotherapy-induced nausea and vomiting with sacituzumab govitecan.⁴⁷ The first sacituzumab govitecan infusion should be administered over 3 hours. It is also recommended that patients be monitored during the entire time of infusion and for ≥30 minutes after completion of the infusion for signs and symptoms of infusion-related reactions.

The prescribing information for sacituzumab govitecan⁴⁷ recommends stopping or discontinuing treatment with sacituzumab govitecan to manage AEs. For grade 4 nonhematologic toxicity of any duration; any grade 3-4 nausea, vomiting, or diarrhea not controlled with antiemetics and antidiarrheal medications; or grade 3-4 nonhematologic toxicity persisting >48 hours despite optimal management, and if at the time of scheduled treatment grade 3-4 toxicity delays the next dose by 2-3 weeks for recovery to grade ≤1, it recommends the following: On first occurrence, lower the dose by 25%, on second occurrence, lower the dose by 50%, and on third occurrence, discontinue treatment. Lastly, for any grade 3-4 nonhematologic or hematologic toxicity that does not recover to grade ≤1 within 3 weeks, we must permanently discontinue treatment.

I am careful with the dose of sacituzumab govitecan. I generally dose reduce in patients with substantial liver metastasis, diminished performance status, a very heavy degree of pretreatment, a heavy burden of cancer, or lower albumin—things like these tell me that my patient is frail. In such individuals, I start with 7.5 mg/kg instead of the recommended starting dose of 10 mg/kg. In robust patients, I do start with 10 mg/kg. At this time, there is no evidence that we need to check UGT1A1 mutation status and dose reduce based on that because there is not a substantial toxicity differential due to the effects of UGT1A1 on SN-38 metabolism. 

Neutropenia associated with sacituzumab govitecan can be managed with either G-CSF or dose reduction.^47,48

In patients with grade 4 neutropenia lasting 7 days or longer, in patients with grade 3-4 febrile neutropenia, or if at time of scheduled treatment grade 3-4 neutropenia delays treatment dosing by 2-3 weeks for recovery to grade ≤1, the prescribing information recommends the following: On first occurrence, lower the dose by 25% and administer G-CSF; on second occurrence, lower the dose by 50% and administer G-CSF; on third occurrence, permanently discontinue treatment and administer G-CSF.⁴⁷ Lastly, if at any time of scheduled treatment a grade 3-4 neutropenia delays dosing by more than 3 weeks for recovery to grade ≤1, we must discontinue treatment and administer G-CSF.

We can use short-acting G-CSF on Days 2-4 and/or long-acting G-CSF on Day 9. I usually will give short-acting G-CSF on Day 5 or 6 prior to Day 8, because I want to be able to deliver that Day 8 dose of sacituzumab govitecan. I do try to maintain the dose intensity of sacituzumab govitecan, and if neutropenia becomes a challenge, I try to use hematopoietic growth factors prior to dose reduction. By contrast, I do tend to dose reduce more for diarrhea. 

If patients have prolonged neutropenia, then that is where you may want to check the UGT181*28 mutation status.⁴⁹ Of course, we are going to want to dose reduce sacituzumab govitecan if patients develop severe, prolonged grade 4 neutropenia or febrile neutropenia. Certainly, for any grade 3 diarrhea or recurrent grade 2 diarrhea, I also would dose reduce in this scenario.