Everolimus Plus Exemestane Treatment in Patients with Metastatic Hormone Receptor‐Positive Breast Cancer Previously Treated with CDK4/6 Inhibitor Therapy

Abstract

Background

The combination of everolimus (EVE) and exemestane (EXE) is approved for the treatment of patients with metastatic hormone receptor‐positive breast cancer (mHRBC) who progress on nonsteroidal aromatase inhibitor (NSAI) therapy. However, none of the patients enrolled in the trial that led to this approval (BOLERO‐2) had previously received CDK4/6 inhibitors (CDK4/6is), which have since become a frontline standard of care for mHRBC. As such, the clinical benefit of EVE plus EXE in patients who have previously received CDK4/6is remains unknown.

Materials and Methods

Adult patients with mHRBC at our institution who progressed on an NSAI plus CDK4/6i or NSAI therapy alone and were treated with at least one cycle of EVE plus EXE between 2012 and 2018 were analyzed. Collected data included patient demographics, treatment history, adverse events, and clinical outcomes. Primary objectives were to compare progression‐free survival (PFS) and overall survival (OS) between patients who received prior NSAI plus CDK4/6i therapy versus an NSAI alone.

Results

Among 43 patients, 17 had prior CDK4/6i exposure. With the exception of de novo metastatic disease, patient and disease characteristics were comparable across treatment cohorts. There was no significant difference in PFS (median, 3.6 vs. 4.2 months) or OS (median, 15.6 vs. 11.3 months) between patients who had received prior CDK4/6is and those who had not, respectively.

Conclusion

Prior exposure to CDK4/6i therapy did not impact survival outcomes for patients with mHRBC taking EVE plus EXE. However, there was a trend toward improved OS in the CDK4/6i cohort that should be evaluated in larger cohorts.

Implications for Practice

The use of CDK4/6 inhibitors in combination with a nonsteroidal aromatase inhibitor has become a standard frontline therapy in metastatic hormone receptor‐positive breast cancer. An approved subsequent line of therapy is everolimus plus exemestane; however, the original data supporting this therapy predated approval of CDK4/6 inhibitors. As such, the clinical benefit of everolimus and exemestane in patients previously treated with a CDK4/6 inhibitor was unknown. This retrospective cohort study offers real‐world data demonstrating prior CDK4/6 inhibitor exposure does not impact survival outcomes for everolimus plus exemestane.

Short abstract

The clinical benefit of everolimus and exemestane in patients previously treated with a CDK4/6 inhibitor is unknown. This study used real‐world data to investigate survival outcomes in such cases.

Introduction

Hormone receptor‐positive (HR+), human epidermal growth factor receptor 2‐negative (HER2−) breast cancer is the most common subtype of breast cancer [1]. The Surveillance, Epidemiology, and End Results Program database estimates that the age‐adjusted rate is 87 new cases per 100,000 women [1]. The 5‐year survival rate is 100% for localized HR+/HER2− disease but just 30.4% in patients with distant disease [1]. For many years, the mainstay of treatment for patients with metastatic HR+ breast cancer (mHRBC) has been sequential single‐agent endocrine therapy, followed by palliative chemotherapy [2]. The last decade witnessed the development of novel targeted therapies for this population, including mammalian target of rapamycin (mTOR) inhibitors and cyclin‐dependent kinase 4/6 inhibitors (CDK 4/6is) [3]. As these new medications are approved and widely adopted as frontline therapies, there is a need to examine real‐world data to determine their impact on the effectiveness of later lines of treatment.

The combination of a nonsteroidal aromatase inhibitor (NSAI) and CDK 4/6i represents an important new treatment option for patients with mHRBC. This is based on the results of the PALOMA‐2, MONALEESA‐2, and MONARCH‐3 clinical trials, which demonstrated that the combination of a CDK 4/6i (palbociclib, ribociclib, or abemaciclib) and an NSAI significantly increased progression‐free survival (PFS) compared with an NSAI alone as frontline therapy for patients with mHRBC [4, 5, 6]. Overall survival (OS) data have not been fully reported to date, but the near doubling in median PFS has led to adoption of NSAI and CDK 4/6i combination as a new standard for frontline treatment of mHRBC.

The combination of the mTOR inhibitor, everolimus (EVE), and the steroidal aromatase inhibitor, exemestane (EXE), is an approved second line therapy for mHRBC. This approval was based on the BOLERO‐2 trial, which demonstrated a significant improvement in PFS with the combination of EVE plus EXE (median, 6.9 months) compared with EXE alone (median, 2.8 months) [7]. The median OS in the EVE plus EXE cohort was greater than in the EXE cohort (31 vs. 26.6 months, respectively) but did not reach statistical significance [8]. It is important to note that other trials investigating mTOR inhibitors have not shown benefit compared with everolimus, which further solidifies everolimus as the only approved mTOR inhibitor in mHRBC. BOLERO‐2 included patients whose disease was refractory to an NSAI (letrozole or anastrozole) but not patients with prior CDK4/6i exposure because this trial was conducted years before the above‐mentioned CDK4/6i trials reported [9]. Therefore, the clinical benefit of EVE plus EXE is unknown among patients who have progressed on combination therapy of an NSAI and a CDK4/6i.

Materials and Methods

Patient Population

This is a retrospective cohort study of patients treated at our institution, the Knight Cancer Institute at Oregon Health and Science University, a comprehensive cancer center. Clinical practice consists of clinics at the main hospital, as well as satellite clinics (Community Hematology and Oncology) providing clinical services to Portland‐metro area. At OHSU medical oncology clinics, we see approximately 750–1,000 new breast cancer cases per year.

Electronic pharmacy records and departmental databases were reviewed to identify patients with mHRBC aged 18 years or older who received everolimus therapy at our institution between 2012 and 2018. Electronic medical records were manually reviewed to confirm that everolimus was taken with exemestane. Patients were then reviewed for inclusion and exclusion criteria. Patients were required to be hormone receptor positive, defined as estrogen receptor (ER) and/or progesterone receptor (PR) expression–positive by immunohistochemistry (IHC) according to College of American Pathology (CAP) guidelines on either the initial biopsy or the biopsy at the time of metastatic diagnosis. Patients were excluded for being HER2‐positive per CAP guidelines (by IHC or in situ hybridization) or for receiving EVE plus EXE for less than 28 days (i.e., less than one cycle). Patients were then reviewed for receipt of a CDK4/6i. Identified patients were divided into two cohorts: prior CDK4/6i therapy and CDK4/6i‐naive (control). Chart review and data collection was done by one author (M.C.) and reviewed by a different author (L.A., Z.M.) to ensure data accuracy. All abstractors were trained in chart reviews. The institutional review board's approval and waiver of informed consent were obtained prior to commencing the chart review.

Data Collection

Extracted data included patient demographics, tumor characteristics, information on prior therapies (i.e., chemotherapy, endocrine therapy, CDK4/6i), sites of metastatic disease, special interest adverse events due to everolimus, duration of EVE plus EXE, reason for EVE plus EXE discontinuation, disease response and progression, and survival data. Both chart notes and medication documentation within the electronic health record were reviewed for receipt of the therapies studied. Date of disease progression was defined as the date of the imaging study that showed progression or the date of the office visit where the provider documented clinical progression. Information on adverse events was collected by review of in office encounters, admissions, and documented telephone calls. Data on adverse events were collected through everolimus treatment and up to 28 days following the final dose. Only specific adverse events of special interest were evaluated: stomatitis, hand foot syndrome, and pneumonitis. Data on clinical outcomes were collected through July 1, 2019.

Statistical Analysis

The primary endpoints for this study consisted of PFS and OS for the two cohorts. Secondary endpoints included duration of EVE plus EXE therapy, EVE‐related special interest adverse event and dose reduction rates, and prophylactic steroid mouthwash effectiveness. The start time for survival analyses was the date of first everolimus dose. Patient‐level treatment cohort (i.e., CDK4/6i vs. control) characteristics were summarized with descriptive statistics, and differences between cohorts were assessed by Fisher's exact test for categorical variables and the Kruskal‐Wallis test for continuous variables. Time‐to‐event outcomes (PFS, OS, time to EVE plus EXE failure) for each treatment cohort were estimated using the Kaplan‐Meier method and compared with the log‐rank test. After checking the proportional hazards assumption by visually inspecting and testing scaled Schoenfeld residuals, Cox regression models were fit to produce unadjusted or covariate‐adjusted hazard ratios and score or Wald test p values for the treatment group effect. Odds ratios and Fisher's exact test p values were computed to assess the effectiveness of dexamethasone mouthwash in preventing stomatitis. All statistical analyses and figures were created with R version 3.6.1, and statistical significance was ascribed to tests (or the variables being tested) with p values <.05.

Results

Forty‐eight patients with mHRBC treated with EVE plus EXE between 2012 and 2018 were identified as candidates for this retrospective study. Five of these patients were excluded from analysis as they received EVE for less than one cycle (i.e., 28 days). Of the 43 included patients, 17 had received treatment with a CDK4/6i, and 26 were CDK4/6i naive. Of the 17 patients exposed to a CDK4/6i prior to EVE plus EXE, 16 were treated with palbociclib and 1 was treated with ribociclib. All patients in this group were treated after U.S. Food and Drug Administration (FDA) approval of these medications except one who received palbociclib on a phase III trial, then transitioned to off‐trial treatment after FDA approval. The median time between CDK4/6i end date and EVE start was 12 days in the CDK4/6i cohort. Four of the 17 patients had greater than 30 days between CDK4/6i and EVE; within this group, two received intervening therapies, one continued NSAI alone, and one did not receive any intervening treatment. Patient‐level characteristics are summarized by treatment cohort in Table 1. The percentage of patients presenting with de novo metastatic disease was the only statistically significant difference between the cohorts (41.2% for CDK4/6i vs. 11.5% for control, p = .034).

Table 1.

Baseline patient and clinical characteristics

Characteristic	CDK4/6i (n = 17)	Control (n = 26)	p value
De novo metastatic disease, n (%)			.034
No	10 (58.8)	23 (88.5)
Yes	7 (41.2)	3 (11.5)
Age at Dx, median (range), yr	55.0 (31.0–78.0)	52.5 (29.0–79.0)	.737
No. prior treatments in metastatic setting, median (range)	1.0 (1.0–8.0)	2.0 (0.0–6.0)	.327
No. prior tx (binary), n (%)			.344
1	9 (52.9)	9 (34.6)
>1	8 (47.1)	17 (65.4)
Brain metastasis (at Met Dx), n (%)			>.999
No	16 (94.1)	25 (96.2)
Yes	1 (5.9)	1 (3.8)
Bone metastasis (at Met Dx), n (%)			>.999
No	3 (17.6)	5 (19.2)
Yes	14 (82.4)	21 (80.8)
Lung metastasis (at Met Dx), n (%)			.113
No	8 (47.1)	19 (73.1)
Yes	9 (52.9)	7 (26.9)
Liver metastasis (at Met Dx), n (%)			>.999
No	12 (70.6)	18 (69.2)
Yes	5 (29.4)	8 (30.8)
Mo from Met Dx to CDK4/6i start, median (range)	3.7 (0.4–105.6)	N/A	N/A
Duration of CDK4/6i, median (range), mo	10.3 (2.8–33.4)	N/A	N/A
Days from CDK4/6i end to EVE start, median (range)	12.0 (1.0–408.0)	N/A	N/A
Mo from Met Dx to EVE start, median (range)	32.2 (3.4–111.8)	17.2 (0.3–114.3)	.139
Duration of EVE treatment, median (range), mo	3.6 (1.0–12.8)	4.5 (0.9–16.3)	.535
Follow‐up time from EVE start, median (range), mo	13.2 (2.5–36.0)	11.3 (2.0–75.4)	.700

Abbreviations: CDK4/6i, cyclin‐dependent kinase 4/6 inhibitor; Dx, diagnosis; EVE, everolimus; Met, metastatic; N/A, not applicable.

The median time from metastatic diagnosis to commencing EVE plus EXE combination therapy was 32.2 months for the CDK4/6i cohort and 17.2 months for the CDK4/6i‐naive (control) cohort (p = .139). Among the 17 patients receiving CDK4/6i alongside an NSAI, the CDK4/6i therapy was started, on average, 19.3 months (median, 3.7 months) after the diagnosis of metastatic disease and administered for a median of 10.3 months. The duration of EVE plus EXE treatment was not significantly different between the two cohorts, with a median (interquartile range) of 3.6 (2.9–5.9) months for CDK4/6i patients versus 4.5 (2.2–9.5) months for control patients. At 6 months after EVE start, 17.6% (95% confidence interval [CI], 6.3%–49.3%) of CDK4/6i patients were still taking EVE plus EXE compared with 42.3% (95% CI, 27.0%–66.3%) in the control cohort. An adverse event led to everolimus discontinuation in nine (20.9%) included patients: four patients (23.5%) in the CDK4/6i versus five patients (19.2%) in the control cohort. Progression led to everolimus discontinuation in 32 patients: 13 (76.5%) in the CDK4/6i cohort and 19 (73.1%) in the control cohort. Other reasons for discontinuation of EVE plus EXE included transition to hospice (two patients in the control cohort) and death while on treatment (one control cohort patient, exact cause unknown).

Nine (20.9%) of the 43 patients required a dose reduction of everolimus (11.8% for CDK4/6i vs 26.9% for control, p = .281). Nineteen (44.2%) of the 43 included patients developed stomatitis, 15 (78.9%) of whom had no documentation of receiving prophylactic dexamethasone mouthwash. Consequently, the odds of experiencing stomatitis during or up to 28 days after everolimus therapy were 55% lower for subjects who preventatively used this steroid treatment (odds ratio [OR], 0.45; 95% confidence interval [CI], 0.08–2.08; p = .324). In the control cohort, one patient developed hand‐foot syndrome and two patients were diagnosed with pneumonitis. No CDK4/6i patients experienced either of these adverse effects.

At the time of data cutoff, all patients had either progressed or died; thus, there were no censored PFS times. Disease progression was determined from imaging in all patients except one in the control group whose progression was based on worsening symptoms and an increase in cancer antigen 27‐29. Median PFS estimates were 3.6 months for the CDK4/6i cohort and 4.2 months for the control cohort (hazard ratio [HR], 1.22; 95% CI, 0.65–2.28, p = .538), as shown in Table 2 and Figure 1. When controlling for patient‐level characteristics with univariable Cox model p values <.100 (i.e. age, bone metastasis, time from metastatic diagnosis to EVE plus EXE start), there was minimal change to the treatment group hazard ratio and p value (Table 2). Of the nine patients who discontinued EVE plus EXE because of an adverse event, three patients went on to receive cancer treatment between discontinuing and progression. Two of these patients started a different endocrine monotherapy, and one continued exemestane alone. Disease response rate was not calculated because of lack of consistent baseline and follow‐up imaging.

Table 2.

Progression‐free survival and overall survival estimates from Cox models

Covariate in model	HR, CDK4/6i vs. control (HR 95% CI)	p value
Progression‐free survival
None	1.22 (0.65–2.28)	.538
Age at diagnosis	1.18 (0.63–2.22)	.603
Bone metastasis	1.16 (0.62–2.19)	.638
Months from Met Dx to EVE	1.11 (0.58–2.11)	.759
Overall survival
None	0.70 (0.35–1.40)	.308
De novo Met disease	0.98 (0.46–2.10)	.964

Abbreviations: CDK4/6i, cyclin‐dependent kinase 4/6 inhibitor; CI, confidence interval; Dx, diagnosis; EVE, everolimus; HR, hazard ratio; Met, metastatic.

Figure 1.

Progression‐free survival by CDK4/6i exposure.Abbreviation: CDK4/6i, CDK4/6 inhibitor.

Deaths were recorded in 37 of the 43 included patients with mHRBC and ranged from 2 months to 4.4 years after the first dose of everolimus, with OS times censored in 1 control patient and 5 CDK4/6i patients. Although median OS was greater for the CDK4/6i group (15.6 months) than for the control group (11.3 months), there was not evidence of a survival difference throughout follow‐up (HR, 0.70; 95% CI, 0.35–1.40, p = .308; Table 2, Fig. 2). In a multivariable model, de novo metastatic disease correlated with longer OS (HR, 0.41; 95% CI, 0.16–1.08; p = .072) whereas there was no effect of CDK4/6i (HR, 0.98; 95% CI, 0.46–2.10, p = 0.964; Table 2).

Figure 2.

Overall survival by CDK4/6i exposure.Abbreviation: CDK4/6i, CDK4/6 inhibitor.

Discussion

Therapy for mHRBC has evolved significantly in the past decade. From the approval of tamoxifen in 1977 through 2012, treatment consisted of sequential endocrine therapies and palliative chemotherapy [2]. Since 2012, targeted therapies, including mTOR inhibitors and CDK4/6is, have become standard of care in combination with endocrine therapies for mHRBC. With the constant evolution in mHRBC therapies, there is a need to examine the real‐world impacts beyond the single line of therapy evaluated in the trials leading to the drug approvals. As new frontline therapies are developed and approved, the subsequent sequential therapies may need to be re‐evaluated. Everolimus plus exemestane is an example of a second‐line therapy that was studied under a frontline therapy that has since changed. Now CDK4/6is have been adopted as an addition to NSAIs for frontline therapy. Here, we evaluated whether this new frontline therapy had any impact on the clinical benefit of EVE plus EXE as a second‐line therapy for mHRBC.

In an exploratory analysis of the PALOMA‐3 study, which evaluated the CDK4/6i palbociclib plus fulvestrant versus fulvestrant alone in patients with mHRBC, investigators analyzed time on subsequent lines of therapy [10]. For patients who received EVE plus EXE following PALOMA‐3, the median duration of EVE plus EXE therapy was 4.3 months (95% CI, 2.5–7.6) for the palbociclib plus fulvestrant group and 5.0 months (95% CI, 2.5–9.4) for the placebo plus fulvestrant group [10]. Additionally, a recent single‐center retrospective analysis of subsequent therapy after palbociclib found EVE plus EXE is the most commonly prescribed endocrine‐based therapy after progression, with n = 12 [11]. The median PFS for EVE plus EXE postpalbociclib in this study was 4.9 months. Our study had similar findings to these two studies. Importantly, our study offers a more specific look at patients previously treated with a CDK4/6i plus an NSAI or NSAI therapy alone in a real‐world setting. We also evaluated EVE plus EXE therapy after any CDK4/6is, including but not exclusive to palbociclib, which is important as there may be different CDK4/6i prescribing patterns across different practices.

Our results showed that CDK4/6i exposure prior to EVE plus EXE therapy did not significantly impact PFS or OS in patients with mHRBC. Our estimated PFS and OS times were lower than what was reported in the BOLERO‐2 trial; however, our patient population was more heavily pretreated than in BOLERO‐2, potentially explaining this finding. Additionally, PFS measurement and comparison to clinical trials from a retrospective review of real‐world data is limited, as there are not controlled imaging follow‐up timelines and no use of RECIST criteria for progression. Of note, median OS in our study was longer, albeit nonsignificantly, in the cohort that had received prior CDK4/6i therapy (median of 15.6 vs. 11.3 months, p = .308) [7] However, when statistically adjusting for de novo metastatic disease, there was virtually no survival difference between the two cohorts.

We also evaluated adverse events (AEs) of special interest related to everolimus therapy between the two cohorts. We found that 44.2% of evaluated patients developed stomatitis; however, 78.9% of these patients had no documentation of prophylactic dexamethasone mouthwash, which was not necessarily standard therapy at that time. This study's high rate of stomatitis is a little under the historical rate of 56% from BOLERO‐2, in which patients were not provided prophylactic dexamethasone mouthwash [7]. Our findings further reinforce the importance and benefit of prophylactic steroid mouthwash to decrease stomatitis rates, as reported in the SWISH trial [12].

Other everolimus related AEs of special interest we evaluated were hand‐foot syndrome and pneumonitis, which occurred in 2.3% and 4.7% of patients, respectively. Overall, 20.9% of all patients required an everolimus dose reduction. Although 9 of the included 43 patients discontinued everolimus because of an adverse event, 4 of the 5 patients who were excluded for receiving less than one cycle of everolimus stopped taking this drug because of an AE. When evaluating these five patients alongside the included set of 43, 27.1% of patients with mHRBC who commenced EVE plus EXE therapy permanently ended everolimus because of an AE. Acknowledging that adverse events are difficult to fully evaluate retrospectively, the rates of special interest AEs, everolimus dose reductions, and AE‐related discontinuations did not significantly differ between the two cohorts. This indicates that prior CDK 4/6i therapy should not preclude use of EVE plus EXE as a safe and effective treatment option.

Our study limitations include being a retrospective chart review with a limited sample size from a single institution. As such, lack of statistical significance for various measures tested across treatment groups could be false negatives because there was no power analysis for this retrospective study. Additionally, the studied therapies were not constrained to specific lines of therapy. Some patients had multiple lines of therapy before CDK4/6i or EVE plus EXE. Ideally, the population would be narrowed to CDK4/6i as frontline therapy and EXE plus EVE as second‐line therapy only. Furthermore, it's important to note that 3 of the 17 patients in the CDK4/6i cohort had some cancer therapy between CDK4/6i and EVE plus EXE. These constraints were not used in our study to allow for a larger sample size but may have impacts on our results. However, in a changing treatment landscape for mHRBC, our study provides real‐world data to support the EVE plus EXE regimen as a safe and effective second line therapy in mHRBC, even in patients with prior CDK 4/6i exposure.

Author Contributions

Conception/design: Madeline M. Cook, Luai Al Rabadi, Michael A. Savin, Zahi I. Mitri

Provision of study material or patients: Madeline M. Cook, Luai Al Rabadi, Megan M. Saraceni

Collection and/or assembly of data: Madeline M. Cook, Luai Al Rabadi

Data analysis and interpretation: Madeline M. Cook, Luai Al Rabadi, Andy J. Kaempf, Zahi I. Mitri

Manuscript writing: Madeline M. Cook, Luai Al Rabadi, Andy J. Kaempf, Megan M. Saraceni, Michael A. Savin, Zahi I. Mitri

Final approval of manuscript: Madeline M. Cook, Luai Al Rabadi, Andy J. Kaempf, Megan M. Saraceni, Michael A. Savin, Zahi I. Mitri

Disclosures

The authors indicated no financial relationships.