Real World Experience of Second-Line Treatment Strategies after Palbociclib and Letrozole: Overall Survival in Metastatic Hormone Receptor-Positive Human Epidermal Growth Factor Receptor 2-Negative Breast Cancer

Simple Summary Second-line treatment strategy after the first-line CDK4/6 inhibitor with aromatase inhibitor is considered by the behavior of hormone receptor positive human epidermal growth factor receptor-2 negative (HR+HER2−) metastatic breast cancer (MBC). Progression free survival 2 was one of the association factors for overall survival of HR+HER2− MBC. Therefore, the second-line treatment strategy was important to improve prognosis in patients with HR+/HER2− MBC. Abstract Background: We analyzed real-world practice of second-line treatment in hormone receptor (HR)+ human epidermal growth factor receptor-2 (HER2)− metastatic breast cancer (MBC) following the first-line CDK4/6 inhibitor with letrozole. In addition, we evaluated the relationship between second-line treatment strategies and survival outcome. Methods: Using the clinical data warehouse, clinical information including MBC diagnosis, treatment and survival outcomes were collected. Results: In total, 305 patients were treated with the first-line palbociclib plus letrozole, and we evaluated 166 patients who were treated with second-line treatment. Of the 166 patients, 28.5% were treated with capecitabine (C), followed by exemestane with everolimus (EE) (27.3%) or cytotoxic chemotherapy other than capecitabine (T) (18.8%) and fulvestrant-based treatment or endocrine monotherapy (F) (12.7%). Eighteen patients (10.9%) were enrolled in clinical trials (CT). With regard to treatment strategies, and the median progression-free survival of second-line treatment in a metastatic setting (PFS2) was 7.4 months with C, 5.2 months with EE, 4.8 months with T, 3.6 months with F, and 3.6 months with CT (p = 0.066). In patients with visceral organ disease progression, C (31.3%) or T(31.3%) was the most common second-line treatment followed by EE (21.9%). Most of the 47 patients with bone metastasis alone were treated with EE (38.2%), followed by C (23.4%) and F (21.3%) (p = 0.008). The median overall survival of second-line treatment in a metastatic setting (OS2) was 42.3 months with C, 35.7 months with F, 30.7 months with EE, and 23.1 months with T. The median OS2 for those in CT was not reached (p = 0.064). ER driven BC, disease progression site and PFS2 were associated with OS and OS2 in HR+HER2− MBC (ps < 0.05). Conclusions: We suggested the second line treatment strategy was important to improve prognosis in patients with HR+/HER2− MBC, especially given the recent standardization of first-line treatment and the many available second-line options.


Introduction
The combination of a cyclin-dependent kinase 4 and 6 (CDK4/6) inhibitor with endocrine therapy (ET) is the current standard of care for patients with hormone receptorpositive (HR+) human epidermal growth factor receptor 2-negative (HER2−) metastatic breast cancer (MBC) which is defined by the expression of either the estrogen receptor (ER) or progesterone receptor (PgR) in at least 1% of tumor cells and the absence of HER2 overexpression or amplification [1,2]. There are three CDK4/6 inhibitors: palbociclib, abemaciclib, and ribociclib. Several clinical trials have demonstrated significantly improved progressionfree survival (PFS) with the combination of a CDK4/6 inhibitor and an aromatase inhibitor (AI) compared to an AI alone as the first-line treatment of patients with HR+HER2− MBC regardless of menopausal status [3][4][5][6]. In terms of overall survival (OS), the combination of an AI with ribociclib or abemaciclib improved survival outcomes compared to AI treatment alone. However, AI plus palbociclib did not have a statistically significant OS benefit [7][8][9][10]. Recent real-world database analysis demonstrated significantly longer OS with palbociclib plus AI compared to AI alone as the first-line treatment for HR+/HER2− MBC, but no OS benefit in the PALOMA-2 clinical trial caused by palbociclib could not be included in category 1A recommendation in the current practice guidelines [2,11,12].
In second-line treatment, endocrine therapy (ET) with or without targeted agents is recommended as the second-line treatment after CDK4/6 inhibitor use, unless the patient is refractory to ET according to the treatment guidelines for HR+HER2− MBC [2,12]. However, despite the recent treatment advancements using the combination of a CDK4/6 inhibitor and an AI, approximately 50% of patients pass away within five years after an MBC diagnosis [7][8][9][10]. Therefore, it is essential to determine the optimal treatment sequence after the use of a CDK4/6 inhibitor in order to improve survival outcomes.
There are other new therapeutic strategies beyond the CDK4/6 inhibitor. For instance, the combined use of a PI3Kα-specific inhibitor (alpelisib) with ET has significantly longer PFS in PIK3CA mutant HR+HER2− MBC as the second-line treatment [13,14]. In addition, trastuzumab deruxtecan (T-Dxd) has a survival benefit in HR+HER2− low MBC included in HR+HER2− MBC [15]. However, genetic tests such as PIK3CA cannot be performed in real-time, and some of these novel therapies are not widely available.
In this study, we analyzed a consecutive retrospective cohort of patients with HR+/HER2− MBC who were treated first-line with palbociclib and letrozole. Our aim was to evaluate real-world long-term survival outcomes of palbociclib with letrozole as a first-line therapy, second-line treatment patterns after palbociclib, factors influencing treatment choice, and the prognosis associated with each selected treatment.

Patients
We collected patient data from the clinical data warehouse (CDW) in Samsung Medical Center (SMC). We selected data from HR+HER2− MBC patients who were treated with palbociclib and letrozole as the first-line treatment in metastatic setting therapy between January 2014 and December 2020. The diagnostic studies for MBC included chest computed tomography (CT), abdomino-pelvic CT, bone scan, or positron emission tomography-CT and brain imaging if indicated. HR and HER2 status examinations were permitted in metastatic biopsies as well as archival tissues. The details of BC pathology analysis were described in a previous study [16].
For analysis of second-line treatments, we excluded patients who had been lost to follow-up after first-line palbociclib with letrozole or the first cycle of second-line treatment.

Statistical Analysis
PFS was defined as the time from initiation of palbociclib with letrozole to disease progression or death from any cause, whichever occurred first. In addition, PFS2 was defined by the time from initiation of the second-line treatment to disease progression or death. The OS was defined as the time from initiation of palbociclib with letrozole to death Cancers 2023, 15, 3431 3 of 15 from any cause, and the OS2 was the time between the start of the second-line treatment and death. PFS and OS were analyzed using the Kaplan-Meier method. Cox proportional hazard regression was used to estimate the hazard ratios and 95% confidence intervals (CIs). Correlations between clinical characteristics and tumor response were analyzed using a two-sided Student's t-test and Fisher's exact test. Two-tailed p-values < 0.05 were considered statistically significant. All statistical analyses were performed using IBM SPSS Statistics, ver. 29 (IBM Co., Armonk, NY, USA).

Patient Baseline Characteristics and Updated Survival Analysis
We included 305 patients with a data cut-off date of 3 April 2023 ( Figure 1). By that date, 181 patients (59%) experienced disease progression after first-line therapy with palbociclib and letrozole. Among these 181 patients, three died and five were lost to followup at that time of disease progression. Among the remaining 173 patients who received second-line treatment for MBC in SMC, seven were lost to follow-up after the first cycle of the second-line treatment. Ultimately, we included 166 patients in the second-line treatment analysis.
progression or death from any cause, whichever occurred first. In addition, PFS2 w fined by the time from initiation of the second-line treatment to disease progress death. The OS was defined as the time from initiation of palbociclib with letrozole to from any cause, and the OS2 was the time between the start of the second-line trea and death. PFS and OS were analyzed using the Kaplan-Meier method. Cox propo hazard regression was used to estimate the hazard ratios and 95% confidence in (CIs). Correlations between clinical characteristics and tumor response were analyz ing a two-sided Student's t-test and Fisher's exact test. Two-tailed p-values < 0.05 considered statistically significant. All statistical analyses were performed using IBM Statistics, ver. 29 (IBM Co., Armonk, NY, USA).

Patient Baseline Characteristics and Updated Survival Analysis
We included 305 patients with a data cut-off date of 3 April 2023 ( Figure 1). B date, 181 patients (59%) experienced disease progression after first-line therapy wi bociclib and letrozole. Among these 181 patients, three died and five were lost to f up at that time of disease progression. Among the remaining 173 patients who re second-line treatment for MBC in SMC, seven were lost to follow-up after the firs of the second-line treatment. Ultimately, we included 166 patients in the second-line ment analysis. The baseline patient characteristics according to disease progression are descri Supplementary Table S1. In this analysis, age and ECOG performance status and d disease were not different between the two groups. However, visceral metastasis, n of metastatic sites, germline BRCA status, and initial CA-15-3 and CEA levels were ent between the two groups; patients who had experienced disease progression for ciclib with letrozole or not, respectively (p < 0.05, respectively).
In terms of survival analysis, 181 cases of disease progression and 62 death observed during a median follow up of 41.7 months (interquartile range [IQR]: 33.6 In this survival analysis, the median PFS was 29.0 months (95% confidence in The baseline patient characteristics according to disease progression are described in Supplementary Table S1. In this analysis, age and ECOG performance status and de novo disease were not different between the two groups. However, visceral metastasis, number of metastatic sites, germline BRCA status, and initial CA-15-3 and CEA levels were different between the two groups; patients who had experienced disease progression for palbociclib with letrozole or not, respectively (p < 0.05, respectively).
The reasons for each second-line treatment choice are described in Table 1. We evaluated the impact of clinical characteristics including disease progression site (Table S3) and response duration of first-line palbociclib with letrozole. The response duration cut off was set at 12 months according to the definition of ER-driven disease in the era of CDK4/6 inhibitors. [17] Before the analysis, we excluded two patients who received intrathecal chemotherapy only. Of the remaining 164 patients, 64 developed disease progression in visceral organs (39.0%), 47 in the bone only (28.7%), and 53 at others (32.3%) ( Table 1). In patients with visceral organ disease progression, capecitabine (31.3%) or cytotoxic chemotherapy (31.3%) was the most common second-line treatments followed by exemestane with everolimus (21.9%). Most of the 47 patients with bone metastasis alone were treated with exemestane with everolimus (38.2%), followed by capecitabine (23.4%) and fulvestrant (21.3%) (p = 0.008). The response duration of palbociclib with letrozole did not influence the choice of the second-line treatment strategy (p = 0.209). Other clinical
The reasons for each second-line treatment choice are described in Table 1. We evaluated the impact of clinical characteristics including disease progression site (Table S3) and response duration of first-line palbociclib with letrozole. The response duration cut off was set at 12 months according to the definition of ER-driven disease in the era of CDK4/6 inhibitors [17]. Before the analysis, we excluded two patients who received intrathecal chemotherapy only. Of the remaining 164 patients, 64 developed disease progression in visceral organs (39.0%), 47 in the bone only (28.7%), and 53 at others (32.3%) ( Table 1). In patients with visceral organ disease progression, capecitabine (31.3%) or cytotoxic chemotherapy (31.3%) was the most common second-line treatments followed by exemestane with everolimus (21.9%). Most of the 47 patients with bone metastasis alone were treated with exemestane with everolimus (38.2%), followed by capecitabine (23.4%) and fulvestrant (21.3%) (p = 0.008). The response duration of palbociclib with letrozole did not influence the choice of the second-line treatment strategy (p = 0.209).
Other clinical factors including eastern cooperative oncology group (ECOG) performance status (p = 0.001), initial visceral metastasis (p = 0.006) and endocrine resistance in adjuvant setting (p = 0.011) were associated with second-line treatment strategies.

Effect of Second-Line Treatment According to Site of Disease Progression
The second-line treatment regimens were decided based on the organ of disease progression after first-line therapy. Therefore, we evaluated the effectiveness of second-line treatment according to the sites of disease progression (Figure 4).  In patients with visceral organ disease progression, capecitabine and cytotoxic chemotherapy (except capecitabine) had 5.53 and 5.17 months in PFS2, respectively, while fulvestrant had 0.90 months. However, the second-line treatment strategies were not associated with PFS2 (p = 0.640) ( Figure 4A) or OS2 (p = 0.583) ( Figure 4B). However, patients treated with fulvestrant or other endocrine alone treatments only had 11.5 months of OS2 compared to 32.3 months with capecitabine and 28.2 months with exemestane and everolimus.
In the 47 patients with bone metastasis alone, capecitabine had superior PFS2 than did other second-line regimens (p = 0.011) ( Figure 4C). The median PFS2 was 15.4 months with capecitabine, 8.33 months with exemestane and everolimus, 6.63 months with endocrine only treatment, and 3.67 months with clinical trial enrollment. The two patients who were treated with cytotoxic chemotherapy (excluding capecitabine) had only 1.4 months of PFS2. The OS2 was not reached in this population. However, OS2 differed according to second-line treatment strategy (p = 0.002) ( Figure 4D).
Fifty-three patients had disease progression in locations other than bone and visceral organs; the second-line regimens in these patients marginally affected their PFS2, although without statistical significance (p = 0.096) ( Figure 4E). In this population, capecitabine also had the longest PFS2 (5.70 months) compared to that of the other regimens. The OS2 did not differ according to the second-line treatment (p = 0.826) ( Figure 4F).

Discussion
This up-to-date survival analysis found that first-line treatment of HR+HER2− MBC patients with palbociclib and letrozole provided a median PFS of 29.0 months. The median OS was not reached by the median follow up of 41.7 months. In addition, the five-year OS rate was 66.5%.
PALOMA-2, the pivotal clinical trial for the first-line CDK4/6 inhibitor with AI in HR+HER2− MBC patients, reported a median PFS of 24.8 months [3]. The Asian subgroup analysis in PALOMA-2 found a consistent effect of palbociclib and a median PFS of 25.7 months [18]. In real world data, the median PFS of palbociclib was 19.8 months [11]. In our prior study, we found a median PFS of 28.7 months, which was consistent with our current results [16]. However, the OS benefit of palbociclib has not been consistent among previous studies. In PALOMA-2 clinical trial, the OS was 53.9 months in those treated with palbociclib plus an AI compared to 51.2 months in those treated with an AI alone (p = 0.338). In contrast, real world data found an OS of 57.8 months in those treated with palbociclib plus an AI compared to 43.5 months in those treated with an AI alone (p < 0.001) [7,11]. In our data, the median OS was not reached, and the five-year OS rate was 66.5%.
Before the era of CDK4/6 inhibitors, non-steroidal AI was considered the first-line treatment for HR+HER2− MBC without visceral crisis [19]. Previous clinical trials, as well as real world data analyses for HR+HER2− MBC, found that AI as a first-line therapy provided 3-4 years of OS [20][21][22][23][24]. Consistently, real world data with palbociclib showed 43.5 months of OS in patients treated with AI as the first-line treatment [11].
Recent treatment advances have forced OS improvement in HR+HER2− MBC. In particular, implementation of CDK4/6 inhibitors has significantly changed PFS and OS in this population compared to those in the era before these medications. A recent cohort study suggested that the median OS of HR+HER2− MBC from 2017-2019 was 38.4 months, and that the death risk decreased by 24% compared to that before 2017 [23]. In addition, this cohort study found that the OS was influenced by the age at MBC diagnosis, metastatic location, and number of metastatic sites. Consistent with prior findings, we found that the OS was affected by visceral metastasis and number of metastatic sites. The initial elevation of the serum CA-15-3 and germline BRCA alteration increased the risk of death. Moreover, we attempted to evaluate the association between tumor response to treatment and OS. In this analysis, initial visceral metastasis and number of metastases still affected the OS, whereas the ER driven BC and disease progression site after treatment with palbociclib and letrozole were added as a factor that affects the OS. However, the effect of germline BRCA status has disappeared. These results suggest an interaction between tumor biology and treatment that dynamically affects patient prognosis, while some clinical characteristics of MBC still impact the OS.
In our cohort, capecitabine was the most commonly used drug for second-line treatment. Although capecitabine is considered cytotoxic chemotherapy, it is an oral antimetabolite that is quite tolerable and active in MBC patients [25,26]. Recent clinical trials with palbociclib and ET used capecitabine as the direct competitor in HR+HER2− MBC patients, and capecitabine achieved good survival outcomes [27,28]. In our study, we preferred capecitabine as the second-line treatment in visceral disease progression or other progression (not including bone-only progression). In addition, capecitabine has superior PFS2 compared to those of other agents in multivariate analysis. Indeed, capecitabine also had longer PFS2 duration rather than that of other agents in bone-only progression.
In bone-only progression, fulvestrant was preferred as the second-line treatment. Fulvestrant is a selective estrogen receptor down regulator (SERD) that inhibits ESR1 activity and proteolytic stability. Therefore, mutant ESR1, which is the most common resistance mechanism for AI treatment in HR+HER2− BC, could be effectively inhibited by fulvestrant [29,30]. However, fulvestrant was rarely effective in HR+HER2− MBC after treatment with a CDK4/6 inhibitor and AI in previous clinical trials. Fulvestrant produced a PFS of 1.98 months in the VERONICA clinical trial and of 3.6 months in the CAPItello-291 trial [31,32]. Consequentially, fulvestrant led to 3.67 months of PFS in this study; however, we observed 6.63 months of PFS in fulvestrant and endocrine treatment only in bone lesion progression. Therefore, second-line treatment should be selected with regard to the sites of BC progression.
ER-driven tumors affected the OS and the OS2 but not the PFS2. ER-driven tumors tend to have a relatively long response duration for endocrine therapy and remain sensitive to ET. Therefore, we suggest that ER-driven tumors have a favorable survival outcome regardless of the PFS2 and second-line treatment strategies. In addition, PFS2 also affected to the OS and the OS2. This meant that PFS2 was also significant prognostic factor regardless of ER-driven tumor or not. Moreover, visceral organ disease progression was also significantly associated with OS and the OS2 as well as initial visceral metastasis. This suggested that the dynamic interaction between tumor itself and treatment strategy would impact to patients' survival. Eventually, second line treatment strategy would be important to treat HR+HER2− MBC patients as well as the first line treatment strategy.
In this real-world study, we rarely performed tissue biopsy after disease progression of CDK4/6 inhibitor. As BC subtype change occurs in 20% of MBC, and the most common mechanism of subtype change is loss of ER and PgR, BC biopsy after disease progression is important to guide second-line treatment strategies. Furthermore, genetic testing can be used to precisely treat patients with progressive MBC. However, tissue biopsy has been rarely performed in real clinic, and therefore, the lack of biopsies and genetic testing was a limitation in our data analysis. We expect that further genome and real time subtype-based treatments would improve survival outcomes in HR+HER2− MBC patients.
Recent advancements in BC treatment have been achieved using genetic information. Targetable genetic alterations are found in~30-40% and~5% of HR+/HER2− BC [33,34]. PIK3CA is the most commonly altered gene and is also targetable. RB1 loss and the ESR1 mutation are associated with resistance to treatment with a CDK4/6 inhibitor and AI [35,36]. Future treatment strategies will include genetic information and targeted agents, including PIK3a inhibitors and next-generation SERDs [13,14,37].
In this study, we chose the second-line treatment based on initial visceral metastasis and the sites of disease progression. Second line treatment affected to the PFS2 but not the OS and the OS2. However, PFS2 was significantly associated with the OS and OS2, and furthermore, our cohort had good OS outcomes compared to those of other previous studies using Palbociclib [7,11]. Therefore, our findings suggest that each line of treatment is important to improve the OS in HR+HER2− MBC.

Conclusions
In conclusion, optimizing the sequence of later-line treatments based on MBC status is important to improve prognosis in patients with HR+/HER2− MBC, especially given the recent standardization of first-line treatment and the many available second-line options.

Institutional Review Board Statement:
This study was reviewed and approved by the Institutional Review Board (IRB) of Samsung Medical Center, Seoul, South Korea (IRB no. 2021-07-131). This study was performed in accordance with the Declaration of Helsinki and the Good Clinical Practice guidelines.

Informed Consent Statement:
The requirement for informed consent was waived due to the use of de-identified medical records with clinical data.

Data Availability Statement:
A data sharing statement provided by the authors is available by request to the corresponding author.

Conflicts of Interest:
None of the authors have any conflict of interest in this study.