Fulvestrant versus capecitabine as maintenance therapy in hormone receptor-positive, HER2-negative metastatic breast cancer after first-line chemotherapy (FAMILY): a multicenter, open-label, randomized, phase 3 trial Authors: Wenjing Wu, Yaping Yang, Haizhu Chen, Yongkui Lu, Yinduo Zeng, Jianli Zhao, Caiwen Du, Ying Lin, Peijian Peng, Mei Huang, Ying Zhang, Quchang Ouyang, Linxiaoxiao Ding, Ziliang Cheng, Yuanyu Wen, Zhiyong Wu, Jinhui Ye, Li Ling, Kun Wang, Ying Lu, Guorong Zou, Jincai Zhong, Hong Wang, Daren Lin, De Zeng, Chengcai Yao, Qianjun Chen, Liqun Zou, Shusen Wang, Yunfang Yu, Zhongyu Yuan, Ying Wang, Herui Yao Published: 2026 ABSTRACT Maintenance therapy is key in prolonging remission and improving quality of life for hormone receptor-positive (HR + ), HER-2 negative (HER2-) metastatic breast cancer (MBC), but optimal strategies remain debated. This multicenter, open-label, randomized phase 3 trial conducted across 22 hospitals in mainland China (ClinicalTrials.gov, NCT04263298; Chinadrugtrials.org.cn, ChiCTR-IIR-17014036) compared the efficacy and safety of fulvestrant versus capecitabine as maintenance therapy in HR + /HER2- MBC patients after achieving objective response or disease control following first-line chemotherapy. The primary end point was investigator-assessed progression-free survival (PFS). A total of 210 patients underwent randomization, with 105 assigned to each group. As of the data cutoff date on March 1, 2025, the median follow-up was 33.6 months (range 1.5–81.1). Median PFS favored fulvestrant compared with capecitabine (17.3 months [95% CI 12.7–23.3] vs. 9.0 months [95% CI 7.5–14.3]; hazard ratio 0.63, 95% CI 0.47–0.99; p = 0.003). Overall survival data were not yet mature. Grade ≥3 adverse events (AEs) occurred in three patients (2.9%) in the fulvestrant group and 11 (10.5%) in the capecitabine group. Treatment discontinuation due to AEs occurred in 0% and 7.6% of patients, respectively. Fulvestrant maintenance therapy significantly improves PFS compared to capecitabine in HR + /HER2- MBC patients achieving objective response or disease control following first-line chemotherapy, with a favorable safety profile. Introduction Hormone receptor-positive (HR + ), human epidermal growth factor receptor-2 negative (HER2-) metastatic breast cancer (MBC) represents the most prevalent subtype of MBC, accounting for approximately two-thirds of cases.1 Endocrine therapy (ET) is generally recommended as the preferred first-line treatment for patients with a low tumor burden. However, chemotherapy is frequently selected for individuals with substantial tumor burden, rapidly progressive disease, or symptomatic visceral involvement requiring rapid tumor shrinkage.1,2 In China, 50-65% of HR + /HER2- MBC patients receive chemotherapy during first-line treatment,3,4 compared with 25-45% of patients in Western countries.5–7 Notably, data from the ESME database reported an even higher rate, exceeding 50%.8,9 Although the RIGHT Choice trial demonstrated that the first-line cyclin-dependent kinase 4/6 (CDK4/6) inhibitor ribociclib plus ET was superior to combination chemotherapy in premenopausal patients with clinically aggressive HR + /HER2- MBC,10 real-world evidence indicates that 13–36% of patients still receive chemotherapy as first-line treatment.11–14 This persistent reliance on chemotherapy is largely driven by disease-related factors as well as limited access to CDK4/6 inhibitors due to cost, reimbursement, and availability constraints in certain regions. Therefore, optimizing strategies to maintain disease control following initial chemotherapy remains a critical clinical challenge. Maintenance therapy aims to prolong disease control while preserving quality of life. Chemotherapy-based maintenance approaches include continuing the original combination regimen, de-escalating to single-agent therapy, or switching to an alternative chemotherapeutic drug.15 Multiple trials have demonstrated that maintenance chemotherapy can extend progression-free survival (PFS), although overall survival (OS) benefits are generally modest.16–20 Notably, the MANTA1 trial found no improvements in PFS or OS with additional courses of paclitaxel in MBC patients after response to first-line anthracycline/paclitaxel chemotherapy.21 Moreover, toxicity remains a concern, highlighting the need for more tolerable maintenance strategies. Capecitabine, an oral chemotherapeutic drug, has emerged as a promising alternative for maintenance therapy in MBC due to its convenience and favorable efficacy. Several studies have consistently demonstrated that capecitabine-based maintenance improves PFS and, in some cases, OS.22–25 Nevertheless, long-term use can lead to adverse effects such as hand-foot syndrome, potentially affecting patients’ quality of life and treatment adherence. Endocrine-based maintenance therapy, in contrast, offers a biologically targeted alternative. Although limited evidence is available, small-scale studies suggest that maintenance ET may provide modest improvements in time to progression or PFS, albeit without significant OS benefits.26–28 Fulvestrant, a selective estrogen receptor antagonist without agonist activity, has shown superior clinical efficacy compared with aromatase inhibitors (AIs) in HR + /HER2- MBC.29–31 In particular, the phase III FALCON trial demonstrated that fulvestrant significantly prolonged PFS compared with anastrozole (16.6 months vs. 13.8 months; hazard ratio 0.797, 95% CI 0.637–0.999, p = 0.0486), supporting its role as a highly effective endocrine therapy option.30 The phase II FANCY trial further highlighted the potential of fulvestrant 500 mg as a maintenance treatment, reporting a clinical benefit rate (CBR) of 76% and a median PFS of 16.1 months in HR + /HER2- MBC patients after first-line chemotherapy, along with an excellent safety profile.32 These findings provide a strong biological and clinical rationale for selecting fulvestrant rather than an AI as endocrine maintenance therapy following chemotherapy-induced disease control. In a previous retrospective analysis, we found that both chemotherapy and ET maintenance significantly prolonged PFS and OS compared with observation, with no significant difference in efficacy between chemotherapy and ET maintenance in HR+ disease.33 These findings provide robust real-world evidence supporting the clinical value of maintenance therapy while also suggesting that ET could be a viable alternative to chemotherapy in this setting. Despite the promising data for both capecitabine and fulvestrant, no head-to-head randomized controlled trials have directly compared these two maintenance strategies in HR + /HER2- MBC. Moreover, while ET combined with CDK4/6 inhibitors is widely accepted as the preferred standard of care internationally, access to these agents remains limited in many regions due to financial and reimbursement barriers. To address this gap, the present randomized, phase III FAMILY trial was designed to compare fulvestrant with capecitabine as maintenance therapy in patients with HR + /HER2- MBC who responded to first-line chemotherapy. This study was not designed to replace CDK4/6 inhibitor-based therapy but rather to define an effective and pragmatic alternative maintenance strategy in real-world clinical settings where such standard treatment is not routinely available. Results Patient characteristics Between May 5, 2018, and November 21, 2023, a total of 210 patients were enrolled and randomized to receive either fulvestrant (n = 105) or capecitabine (n = 105) (Fig. 1). The median ages were 49.1 years (range 29.0–70.0) in the fulvestrant group and 48.1 years (range 23.6–71.8) in the capecitabine group. The majority of patients in both groups had estrogen receptor (ER)-positive (ER + ) and progesterone-positive (PR + ) tumors, accounting for 81.9% in the fulvestrant group and 87.6% in the capecitabine group. Previous exposure to ET in the adjuvant setting was reported in 43 (41.0%) patients in the fulvestrant group and 44 (41.9%) patients in the capecitabine group. Clinically aggressive disease occurred in 62 (59.0%) patients in the fulvestrant group and 61 (58.1%) patients in the capecitabine group. Visceral metastases were present in 57 (54.3%) patients in the fulvestrant group and 68 (64.8%) in the capecitabine group. Approximately 30% of patients had been administered first-line capecitabine-based chemotherapy. Baseline characteristics were well balanced across the two treatment groups (Table 1).Fig. 1Recruitment, randomization, and flow of trial participantsTable 1Baseline patient characteristicsCharacteristicsFulvestrant (n = 105)Capecitabine (n = 105)Age, yearsMean ( ± SD)50.6 ± 9.549.7 ± 9.6Median (Range)49.1 (29.0–70.0)48.1 (23.6–71.8)Menopausal statusPremenopausal49 (46.7)57 (54.3)Postmenopausal56 (53.3)48 (45.7)ER and PR statusER + /PR+86 (81.9)92 (87.6)ER + /PR−19 (18.1)13 (12.4)Ki67 index≥20%79 (75.2)74 (70.4)<20%26 (24.8)31 (29.6)Histological gradeGrade 11 (1.0)0 (0.0)Grade 233 (31.4)41 (39.0)Grade 318 (17.1)13 (12.4)Unknown53 (50.5)51 (48.6)Previous ETNo62 (59.0)61 (58.1)Yes43 (41.0)44 (41.9)AI9 (8.6)11 (10.5)SERM24 (22.9)22 (21.0)SERM followed by AI2 (1.9)1 (1.0)SERM/AI + OFS8 (7.6)10 (9.5)ET sensitivityaSensitive71 (67.6)69 (65.7)Secondary resistance34 (32.4)36 (34.3)Stage IV de novoYes55 (52.4)52 (49.5)No50 (47.6)53 (50.5)Clinically aggressive diseasebYes62 (59.0)61 (58.1)No43 (41.0)44 (41.9)Surgery for primary breast lesioncYes14 (25.5)16 (30.8)No41 (74.5)36 (69.2)Visceral metastasesYes57 (54.3)68 (64.8)No48 (45.7)37 (35.2)First-line chemotherapyMonotherapy16 (15.2)22 (21.0)Taxane16 (15.2)22 (21.0)Combined chemotherapy89 (84.8)83 (79.0)TX31 (29.5)32 (30.5)AC → T12 (11.4)12 (11.4)AT11 (10.5)9 (8.6)TCb9 (8.5)7 (6.7)TC6 (5.7)6 (5.7)Othersd20 (19.0)17 (16.2)Capecitabine-based chemotherapyYes34 (32.4)34 (32.4)No71 (67.6)71 (67.6)Response to first-line chemotherapyCR1 (1.0)1 (1.0)PR48 (45.7)39 (37.1)SD56 (53.3)65 (61.9)Data are presented as n (%) unless stated otherwiseSD standard deviation, ER estrogen receptor, PR progesterone receptor, ET endocrine therapy, AI aromatase inhibitor, SERM selective estrogen receptor modulator, OFS ovarian function suppression, CR complete response, PR partial response, SD stable disease, T taxane, X capecitabine, A anthracycline, C cyclophosphamide, Cb platinumaET-sensitive patients were defined as those who had de novo metastatic disease and received no prior ET or those who had a disease-free interval >12 months from the completion of neoadjuvant or adjuvant ET. Patients with secondary ET resistance were defined as those whose disease relapsed during adjuvant ET but after the first 2 years or relapsed within 12 months of completing adjuvant ETbClinically aggressive disease was defined as symptomatic visceral metastases, rapid disease progression or impending visceral compromise, or severely symptomatic non-visceral disease, per investigator assessmentcThe analysis was restricted to patients diagnosed with stage IV de novo breast cancerdOther chemotherapy regimens included gemcitabine plus a taxane, AC, TAC, gemcitabine plus platinum, vinorelbine plus capecitabine, and other combinations Recruitment, randomization, and flow of trial participants Table 1Baseline patient characteristicsCharacteristicsFulvestrant (n = 105)Capecitabine (n = 105)Age, yearsMean ( ± SD)50.6 ± 9.549.7 ± 9.6Median (Range)49.1 (29.0–70.0)48.1 (23.6–71.8)Menopausal statusPremenopausal49 (46.7)57 (54.3)Postmenopausal56 (53.3)48 (45.7)ER and PR statusER + /PR+86 (81.9)92 (87.6)ER + /PR−19 (18.1)13 (12.4)Ki67 index≥20%79 (75.2)74 (70.4)<20%26 (24.8)31 (29.6)Histological gradeGrade 11 (1.0)0 (0.0)Grade 233 (31.4)41 (39.0)Grade 318 (17.1)13 (12.4)Unknown53 (50.5)51 (48.6)Previous ETNo62 (59.0)61 (58.1)Yes43 (41.0)44 (41.9)AI9 (8.6)11 (10.5)SERM24 (22.9)22 (21.0)SERM followed by AI2 (1.9)1 (1.0)SERM/AI + OFS8 (7.6)10 (9.5)ET sensitivityaSensitive71 (67.6)69 (65.7)Secondary resistance34 (32.4)36 (34.3)Stage IV de novoYes55 (52.4)52 (49.5)No50 (47.6)53 (50.5)Clinically aggressive diseasebYes62 (59.0)61 (58.1)No43 (41.0)44 (41.9)Surgery for primary breast lesioncYes14 (25.5)16 (30.8)No41 (74.5)36 (69.2)Visceral metastasesYes57 (54.3)68 (64.8)No48 (45.7)37 (35.2)First-line chemotherapyMonotherapy16 (15.2)22 (21.0)Taxane16 (15.2)22 (21.0)Combined chemotherapy89 (84.8)83 (79.0)TX31 (29.5)32 (30.5)AC → T12 (11.4)12 (11.4)AT11 (10.5)9 (8.6)TCb9 (8.5)7 (6.7)TC6 (5.7)6 (5.7)Othersd20 (19.0)17 (16.2)Capecitabine-based chemotherapyYes34 (32.4)34 (32.4)No71 (67.6)71 (67.6)Response to first-line chemotherapyCR1 (1.0)1 (1.0)PR48 (45.7)39 (37.1)SD56 (53.3)65 (61.9)Data are presented as n (%) unless stated otherwiseSD standard deviation, ER estrogen receptor, PR progesterone receptor, ET endocrine therapy, AI aromatase inhibitor, SERM selective estrogen receptor modulator, OFS ovarian function suppression, CR complete response, PR partial response, SD stable disease, T taxane, X capecitabine, A anthracycline, C cyclophosphamide, Cb platinumaET-sensitive patients were defined as those who had de novo metastatic disease and received no prior ET or those who had a disease-free interval >12 months from the completion of neoadjuvant or adjuvant ET. Patients with secondary ET resistance were defined as those whose disease relapsed during adjuvant ET but after the first 2 years or relapsed within 12 months of completing adjuvant ETbClinically aggressive disease was defined as symptomatic visceral metastases, rapid disease progression or impending visceral compromise, or severely symptomatic non-visceral disease, per investigator assessmentcThe analysis was restricted to patients diagnosed with stage IV de novo breast cancerdOther chemotherapy regimens included gemcitabine plus a taxane, AC, TAC, gemcitabine plus platinum, vinorelbine plus capecitabine, and other combinations Baseline patient characteristics Data are presented as n (%) unless stated otherwise SD standard deviation, ER estrogen receptor, PR progesterone receptor, ET endocrine therapy, AI aromatase inhibitor, SERM selective estrogen receptor modulator, OFS ovarian function suppression, CR complete response, PR partial response, SD stable disease, T taxane, X capecitabine, A anthracycline, C cyclophosphamide, Cb platinum aET-sensitive patients were defined as those who had de novo metastatic disease and received no prior ET or those who had a disease-free interval >12 months from the completion of neoadjuvant or adjuvant ET. Patients with secondary ET resistance were defined as those whose disease relapsed during adjuvant ET but after the first 2 years or relapsed within 12 months of completing adjuvant ET bClinically aggressive disease was defined as symptomatic visceral metastases, rapid disease progression or impending visceral compromise, or severely symptomatic non-visceral disease, per investigator assessment cThe analysis was restricted to patients diagnosed with stage IV de novo breast cancer dOther chemotherapy regimens included gemcitabine plus a taxane, AC, TAC, gemcitabine plus platinum, vinorelbine plus capecitabine, and other combinations Efficacy The efficacy analysis was conducted on the full analysis set (FAS). As of the data cutoff date on March 1, 2025, the median follow-up duration was 33.6 months (range 1.5–81.1). Based on the FAS, PFS events occurred in 170 patients, with 80 (76.2%) in the fulvestrant group and 90 (85.7%) in the capecitabine group. PFS favored fulvestrant over capecitabine (median PFS: 17.3 months [95% CI 12.7–23.3] vs. 9.0 months [95% CI 7.5–14.3]; hazard ratio 0.63, 95% CI 0.47–0.99; p = 0.003) (Fig. 2). The median total PFS (tPFS) was 22.2 months (95% CI 19.2–26.8) versus 14.4 months (95% CI 12.0-19.8) with fulvestrant and capecitabine (hazard ratio 0.63, 95% CI 0.46–0.85; p = 0.003), respectively (Fig. 3a). Among patients with clinically aggressive diseases, the median PFS was 16.2 months (95% CI 9.5–21.7) with fulvestrant versus 8.4 months (95% CI 6.6–14.1) with capecitabine (hazard ratio 0.67, 95% CI 0.45–0.98; p = 0.04) (Supplementary Fig. 1a); the median tPFS was 21.1 months (95% CI 15.5–26.5) and 13.6 months (95% CI 11.4–19.7), respectively (hazard ratio 0.66, 95% CI 0.45–0.97; p = 0.034) (Supplementary Fig. 2a).Fig. 2Progression-free survival in the full analysis set. Note: The short vertical lines represent censored dataFig. 3Total progression-free survival (tPFS) and total overall survival (tOS) in the full analysis set. a tPFS; b tOS. The short vertical lines represent censored data. Note: tPFS was defined as the time from initiation of first-line chemotherapy to disease progression, death, or the latest follow-up. tOS was defined as the time from initiation of first-line chemotherapy to death from any cause or the latest follow-up Progression-free survival in the full analysis set. Note: The short vertical lines represent censored data Total progression-free survival (tPFS) and total overall survival (tOS) in the full analysis set. a tPFS; b tOS. The short vertical lines represent censored data. Note: tPFS was defined as the time from initiation of first-line chemotherapy to disease progression, death, or the latest follow-up. tOS was defined as the time from initiation of first-line chemotherapy to death from any cause or the latest follow-up By the data cutoff date, OS data were immature, with 33 (31.4%) deaths in the fulvestrant group and 38 (36.2%) deaths in the capecitabine group. The median OS was 74.2 months (95% CI 67.9-not reached [NR]) for the fulvestrant group, compared to 71.7 months (95% CI, 53.2-NR) for the capecitabine group (hazard ratio 0.84, 95% CI 0.53–1.35; p = 0.47) (Supplementary Fig. 3). The median total OS (tOS) was 79.4 months (95% CI 71.7-NR) with fulvestrant versus 77.6 (95% CI 56.8-NR) with capecitabine (hazard ratio 0.84, 95% CI 0.52–1.33; p = 0.45) (Fig. 3b). In patients with clinically aggressive disease, fulvestrant yielded a median OS of 74.2 months (95% CI 46.0-NR) versus 50.0 months (95% CI 33.7-NR) with capecitabine (HR 0.69; 95% CI, 0.39–1.19; p = 0.18) (Supplementary Fig. 1b); the median tOS was 79.4 months (95% CI 51.5-NR) in the fulvestrant group and 54.4 months (95% CI 37.4-NR) in the capecitabine group (hazard ratio 0.67, 95% CI 0.38-1.17; p = 0.162) (Supplementary Fig. 2b). Table 2 presents the proportions of patients achieving an objective response and clinical benefit. Overall, 12 of 105 (11.4%; 95% CI 5.2–17.6%) patients in the fulvestrant group achieved an objective response compared to 8 of 105 (7.6%; 95% CI 2.5–12.8%) patients in the capecitabine group. Clinical benefit was recorded in 101 of 105 (96.2%; 95% CI 92.5–99.9%) patients in the fulvestrant group and 92 of 105 (87.6%; 95% CI 81.2–94.0%) patients in the capecitabine group.Table 2Summary of efficacy end points in the FASNo. (%) of patientsp valueFulvestrant (n = 105)Capecitabine (n = 105)CR0 (0.0)0 (0.0)PR12 (11.4)8 (7.6)SD ( ≥ 24 weeks)89 (84.8)84 (80.0)PD4 (3.8)13 (12.4)ORR12 (11.4)8 (7.6)0.34795%CI5.2–17.62.5-12.8CBR101 (96.2)92 (87.6)0.02395%CI92.5–99.981.2–94.0PFS0.003 (hazard ratio 0.63, 95%CI 0.47–0.99)Number of events8090Median, months (95%CI)17.3 (12.7–23.3)9.0 (7.5–14.3)OS0.47 (hazard ratio 0.84, 95%CI 0.53–1.35)Number of events3338Median, months (95%CI)74.2 (67.9-NR)71.7 (53.2-NR)FAS full analysis set, CR complete response, PR partial response, SD stable disease, PD progressive disease, ORR objective response rate, CBR clinical benefit rate, PFS progression-free survival, CI confidence interval Table 2Summary of efficacy end points in the FASNo. (%) of patientsp valueFulvestrant (n = 105)Capecitabine (n = 105)CR0 (0.0)0 (0.0)PR12 (11.4)8 (7.6)SD ( ≥ 24 weeks)89 (84.8)84 (80.0)PD4 (3.8)13 (12.4)ORR12 (11.4)8 (7.6)0.34795%CI5.2–17.62.5-12.8CBR101 (96.2)92 (87.6)0.02395%CI92.5–99.981.2–94.0PFS0.003 (hazard ratio 0.63, 95%CI 0.47–0.99)Number of events8090Median, months (95%CI)17.3 (12.7–23.3)9.0 (7.5–14.3)OS0.47 (hazard ratio 0.84, 95%CI 0.53–1.35)Number of events3338Median, months (95%CI)74.2 (67.9-NR)71.7 (53.2-NR)FAS full analysis set, CR complete response, PR partial response, SD stable disease, PD progressive disease, ORR objective response rate, CBR clinical benefit rate, PFS progression-free survival, CI confidence interval Summary of efficacy end points in the FAS FAS full analysis set, CR complete response, PR partial response, SD stable disease, PD progressive disease, ORR objective response rate, CBR clinical benefit rate, PFS progression-free survival, CI confidence interval Subgroup analysis The subgroup analysis confirmed a consistent benefit in PFS from fulvestrant maintenance therapy across the majority of subgroups (Fig. 4). Significant improvements in PFS were observed in the fulvestrant group among younger patients ( ≤ 50 years), premenopausal patients, those with PR-positive tumors or Ki67 < 20%. Additionally, other subgroups, including patients without prior ET, those with secondary resistance to prior ET, individuals with de novo stage IV disease, patients with clinically aggressive disease, and those with visceral metastases, also showed a benefit from fulvestrant maintenance therapy. However, fulvestrant did not demonstrate a clear advantage in patients previously treated with capecitabine-based regimens.Fig. 4Subgroup analysis of progression-free survival. PR, progesterone receptor; ET, endocrine therapy; CR, complete response; PR, partial response; SD, stable disease Subgroup analysis of progression-free survival. PR, progesterone receptor; ET, endocrine therapy; CR, complete response; PR, partial response; SD, stable disease Safety All randomized patients were included in the safety analysis set (SS). At the time of data cutoff, 21 patients in the fulvestrant group and 12 patients in the capecitabine group remained on treatment. The median duration of exposure was 15.9 months (range 2.5-69.0) for fulvestrant and 8.7 months (range 1.6-63.7) for capecitabine. Overall, 36.2% (38/105) and 45.7% (48/105) of patients in the fulvestrant and capecitabine groups, respectively, experienced at least one adverse event (AE). Dose reductions due to AEs occurred in six (5.7%) patients in the capecitabine group. Treatment termination because of AEs was reported in 0 patients and eight (7.6%) patients in the fulvestrant and capecitabine groups, respectively. In both treatment groups, disease progression was the leading cause of permanent discontinuation of the study treatment, occurring in 70.5% (74/105) of patients in the fulvestrant group and 81.9% (86/105) in the capecitabine group. Discontinuation of study treatment due to AEs occurred in one patient (1.0%) in the capecitabine group, while no such cases were reported with fulvestrant. Grade 3-4 AEs were reported in 3/105 (2.9%) and 11/105 (10.5%) patients treated with fulvestrant and capecitabine, respectively. No grade 5 AEs were reported. The most common AE was hand-foot syndrome (45/105, 42.9%) in the capecitabine group (Table 3). Most AEs were grade 1 or grade 2. The capecitabine group showed numerically higher rates of hand-foot syndrome (42.9% vs. 2.9%), elevated alanine transaminase (ALT) or aspartate transaminase (AST) (23.8% vs. 8.6%), decreased white blood cells (WBCs) (19.0% vs. 9.5%), and anemia (14.3% vs. 1.0%) than the fulvestrant group. However, injection-site pain occurred more frequently in the fulvestrant group (16.2% vs. 0%).Table 3Adverse events in the safety analysis setFulvestrant (n = 105)Capecitabine (n = 105)Any gradeGrade 1–2Grade 3 or 4Any gradeGrade 1–2Grade 3 or 4Injection-site pain17 (16.2)15 (14.3)2 (1.9)000Neutrophil count decreased15 (14.3)15 (14.3)020 (19.0)19 (18.1)1 (1.0)γ-GGT increased11 (10.5)10 (9.5)1 (1.0)18 (17.1)18 (17.1)0WBC decreased10 (9.5)10 (9.5)020 (19.0)20 (19.0)0Hyperlipidemia10 (9.5)10 (9.5)08 (7.6)7 (6.7)1 (1.0)Bilirubin increased10 (9.5)10 (9.5)04 (3.8)4 (3.8)0ALT or AST increased9 (8.6)9 (8.6)025 (23.8)25 (23.8)0Platelet count decreased8 (7.6)8 (7.6)015 (14.3)15 (14.3)0Acroanesthesia7 (6.7)7 (6.7)09 (8.6)9 (8.6)0Insomnia6 (5.7)6 (5.7)07 (6.7)7 (6.7)0Hot flush6 (5.7)6 (5.7)03 (2.9)3 (2.9)0Asthenia5 (4.8)5 (4.8)011 (10.5)11 (10.5)0Mucositis oral5 (4.8)5 (4.8)05 (4.8)5 (4.8)0Fatigue4 (3.8)4 (3.8)09 (8.6)9 (8.6)0Weight gain4 (3.8)4 (3.8)09 (8.6)9 (8.6)0Muscle pain4 (3.8)4 (3.8)04 (3.8)4 (3.8)0Rash4 (3.8)4 (3.8)08 (7.6)8 (7.6)0Hand-foot syndrome3 (2.9)3 (2.9)045 (42.9)37 (35.2)8 (7.6)Arthralgia or joint disorders3 (2.9)3 (2.9)08 (7.6)8 (7.6)0Bone pain3 (2.9)3 (2.9)011 (10.5)11 (10.5)0Diarrhea2 (1.9)2 (1.9)03 (2.9)3 (2.9)0Abdominal pain2 (1.9)2 (1.9)02 (1.9)2 (1.9)0Decreased appetite2 (1.9)2 (1.9)05 (4.8)5 (4.8)0Pulmonary infection2 (1.9)2 (1.9)02 (1.9)2 (1.9)0Anemia1 (1.0)1 (1.0)015 (14.3)15 (14.3)0Nausea or vomiting1 (1.0)1 (1.0)08 (7.6)8 (7.6)0Headache or dizziness1 (1.0)1 (1.0)09 (8.6)9 (8.6)0Cough1 (1.0)1 (1.0)01 (1.0)1 (1.0)0ALP increased0002 (1.9)2 (1.9)0Vaginitis0001 (1.0)1 (1.0)0Data are presented as n (%) unless stated otherwiseγ-GGT γ-Glutamyltransferase, WBC white blood cell, ALT alanine aminotransferase, AST aspartate aminotransferase, ALP alkaline phosphatase Table 3Adverse events in the safety analysis setFulvestrant (n = 105)Capecitabine (n = 105)Any gradeGrade 1–2Grade 3 or 4Any gradeGrade 1–2Grade 3 or 4Injection-site pain17 (16.2)15 (14.3)2 (1.9)000Neutrophil count decreased15 (14.3)15 (14.3)020 (19.0)19 (18.1)1 (1.0)γ-GGT increased11 (10.5)10 (9.5)1 (1.0)18 (17.1)18 (17.1)0WBC decreased10 (9.5)10 (9.5)020 (19.0)20 (19.0)0Hyperlipidemia10 (9.5)10 (9.5)08 (7.6)7 (6.7)1 (1.0)Bilirubin increased10 (9.5)10 (9.5)04 (3.8)4 (3.8)0ALT or AST increased9 (8.6)9 (8.6)025 (23.8)25 (23.8)0Platelet count decreased8 (7.6)8 (7.6)015 (14.3)15 (14.3)0Acroanesthesia7 (6.7)7 (6.7)09 (8.6)9 (8.6)0Insomnia6 (5.7)6 (5.7)07 (6.7)7 (6.7)0Hot flush6 (5.7)6 (5.7)03 (2.9)3 (2.9)0Asthenia5 (4.8)5 (4.8)011 (10.5)11 (10.5)0Mucositis oral5 (4.8)5 (4.8)05 (4.8)5 (4.8)0Fatigue4 (3.8)4 (3.8)09 (8.6)9 (8.6)0Weight gain4 (3.8)4 (3.8)09 (8.6)9 (8.6)0Muscle pain4 (3.8)4 (3.8)04 (3.8)4 (3.8)0Rash4 (3.8)4 (3.8)08 (7.6)8 (7.6)0Hand-foot syndrome3 (2.9)3 (2.9)045 (42.9)37 (35.2)8 (7.6)Arthralgia or joint disorders3 (2.9)3 (2.9)08 (7.6)8 (7.6)0Bone pain3 (2.9)3 (2.9)011 (10.5)11 (10.5)0Diarrhea2 (1.9)2 (1.9)03 (2.9)3 (2.9)0Abdominal pain2 (1.9)2 (1.9)02 (1.9)2 (1.9)0Decreased appetite2 (1.9)2 (1.9)05 (4.8)5 (4.8)0Pulmonary infection2 (1.9)2 (1.9)02 (1.9)2 (1.9)0Anemia1 (1.0)1 (1.0)015 (14.3)15 (14.3)0Nausea or vomiting1 (1.0)1 (1.0)08 (7.6)8 (7.6)0Headache or dizziness1 (1.0)1 (1.0)09 (8.6)9 (8.6)0Cough1 (1.0)1 (1.0)01 (1.0)1 (1.0)0ALP increased0002 (1.9)2 (1.9)0Vaginitis0001 (1.0)1 (1.0)0Data are presented as n (%) unless stated otherwiseγ-GGT γ-Glutamyltransferase, WBC white blood cell, ALT alanine aminotransferase, AST aspartate aminotransferase, ALP alkaline phosphatase Adverse events in the safety analysis set Data are presented as n (%) unless stated otherwise γ-GGT γ-Glutamyltransferase, WBC white blood cell, ALT alanine aminotransferase, AST aspartate aminotransferase, ALP alkaline phosphatase Discussion The phase III FAMILY trial demonstrated a clinically meaningful and statistically significant improvement in PFS with fulvestrant maintenance therapy compared to capecitabine maintenance therapy in patients with HR + /HER2- MBC who achieved an objective response or disease control following first-line chemotherapy. The median PFS was 17.3 months (95% CI 12.7–23.3) for the fulvestrant group versus 9.0 months (95% CI 7.5–14.3) for the capecitabine group (hazard ratio 0.63, 95% CI 0.47–0.99, p = 0.003). The PFS benefit of fulvestrant maintenance therapy was consistent across most analyzed subgroups, further validating its clinical efficacy. Importantly, the benefit extended to tPFS, which reached 22.2 months with fulvestrant versus 14.4 months with capecitabine. Notably, nearly 60% of patients in this study presented with aggressive disease features, and in this subgroup, fulvestrant yielded superior PFS and tPFS compared to capecitabine. Additionally, patients in the fulvestrant group experienced numerically fewer AEs, underscoring its favorable safety profile. To our knowledge, this study for the first time directly compares fulvestrant with capecitabine as maintenance therapies in this specific patient population, providing a foundation for further research and clinical application. The optimal sequencing of maintenance ET and chemotherapy in HR + /HER2- MBC remains a subject of debate. Our previous study found that no notable difference was observed in PFS and OS between chemotherapy and ET maintenance in this population.33 The phase II FANCY study suggested that fulvestrant 500 mg could be a promising maintenance therapy for HR + /HER2- MBC patients who exhibited no disease progression after first-line chemotherapy.32 Building upon these preliminary results, the FAMILY trial, for the first time, directly compared the efficacy of fulvestrant with capecitabine as a maintenance therapy after first-line chemotherapy. Encouragingly, this trial achieved its primary end point, with a clinically meaningful PFS improvement of 8.3 months in the fulvestrant group. Together, these findings highlight the growing role of endocrine maintenance therapy in improving outcomes for patients with HR + /HER2- MBC. Several biological mechanisms may contribute to the observed efficacy of fulvestrant as maintenance therapy following chemotherapy compared with capecitabine. Fulvestrant functions as a selective estrogen receptor degrader that binds to the ER and promotes its proteasomal degradation, resulting in more complete and sustained inhibition of ER signaling.34 This mechanism may be particularly relevant after chemotherapy, as cytotoxic treatment can alter ER signaling dynamics and tumor biology. In addition, chemotherapy-induced tumor debulking and potential modulation of the tumor microenvironment may reduce the dominance of resistant subclones, thereby creating conditions under which residual hormone-dependent tumor cells remain susceptible to endocrine suppression. In this context, ER degradation by fulvestrant may more effectively suppress residual endocrine-responsive disease, leading to durable disease control in the maintenance setting. The biological mechanisms underlying these observations remain incompletely understood and warrant further investigation. The RIGHT Choice study reaffirmed that first-line CDK4/6 inhibitors combined with ET are superior to chemotherapy in clinically aggressive HR + /HER2- ABC.10 However, in many parts of the world, particularly in developing countries, access to CDK4/6 inhibitors remains limited due to cost, reimbursement barriers, or drug availability. Against this backdrop, the present FAMILY trial addresses an important clinical gap. For patients achieving objective response or stable disease after a short course of chemotherapy (4-8 cycles), timely transition to endocrine maintenance, such as fulvestrant, offers a rational approach that balances efficacy, tolerability, and cost-effectiveness. In patients with clinically aggressive disease, this sequential strategy was associated with a clinically meaningful duration of disease control, as reflected by a median tPFS of 21.1 months. However, given differences in study design and patient populations, cross-trial comparisons should be interpreted with caution. Treatment tolerability remains an important consideration for clinicians when selecting maintenance strategies. In this study, AEs in the fulvestrant group were consistent with its established safety profiles, with no unexpected AEs or new safety concerns reported.29–31 Similarly, AEs with capecitabine in this study were in line with previous data22,35,36. As expected, the overall incidence of AEs appeared to be lower for fulvestrant compared to capecitabine. Notably, in the fulvestrant group, no patients discontinued treatment due to AEs, whereas eight patients (7.6%) in the capecitabine group discontinued treatment, supporting the favorable tolerability of fulvestrant. Capecitabine was associated with a higher frequency of toxicities, particularly hand-foot syndrome, which could adversely impact quality of life during prolonged treatment. Previous studies have reported a higher susceptibility to fluoropyrimidine-related toxicities, including hand-foot syndrome, in Asian populations, which may be relevant when considering the tolerability of prolonged capecitabine exposure.36–39 In contrast, fulvestrant therapy was better tolerated. The incidence of injection-site pain observed with fulvestrant in this study was numerically higher than that reported in the FIRST study (16.2% and 5.9%, respectively).29 However, the incidence of arthralgia or joint disorders seemed to be numerically lower than that reported in previous studies, such as the FIRST study and the FALCON study (2.9%, 9.9%, and 16.7%, respectively).29,30 These findings underscore the excellent safety profile of fulvestrant, supporting its use as a maintenance therapy aimed at delaying relapse while preserving quality of life. Alternative maintenance strategies for HR + /HER2- MBC are currently under active investigation, including the combination of ET with targeted therapies. Increasing attention has been given to the potential role of CDK4/6 inhibitors combined with ET as a maintenance strategy following first-line chemotherapy. The phase II AMICA trial reported encouraging efficacy and safety of the CDK4/6 inhibitor ribociclib plus ET in patients with HR + /HER2- MBC who achieved at least stable disease following first-line chemotherapy.40 Additionally, an ongoing phase II trial (NCT05979220) is currently evaluating the efficacy and safety of another CDK4/6 inhibitor, dalpiciclib, combined with letrozole, in this setting. Given the established efficacy and manageable safety profile of CDK4/6 inhibitors combined with fulvestrant in HR + /HER2- MBC,41–44 future studies may also consider exploring this combination as a potential maintenance strategy after first-line chemotherapy. Our findings have important implications for real-world clinical practice. In many regions, chemotherapy remains the default first-line choice for patients with high tumor burden or visceral crisis. However, indefinite continuation of chemotherapy is not always necessary and may increase the risk of cumulative toxicity. The FAMILY trial demonstrates that, after disease control with short-course chemotherapy, switching to fulvestrant maintenance can achieve durable disease control, especially in clinically aggressive disease, while reducing toxicity and avoiding the financial and logistical barriers associated with immediate CDK4/6 inhibitor use. This strategy aligns well with current treatment goals in MBC and provides a practical, accessible, and cost-effective alternative for first-line care, especially in regions where CDK4/6 inhibitors are not routinely available or reimbursed. Several limitations should be acknowledged. First, while the observed improvements in PFS are statistically and clinically significant, the OS data remain immature, limiting conclusions about long-term survival benefits. Continued follow-up is ongoing, and updated OS analyses will be performed at future prespecified data cutoffs to determine whether the observed PFS benefit translates into a survival advantage. Second, differences between the trial population and real-world clinical practice must be considered. Most patients in current practice receive ET plus a CDK4/6 inhibitor as initial treatment. Since this trial was initiated before the first approval of CDK4/6 inhibitors for MBC in China, patients pretreated with these agents were not included. Therefore, the findings of this study are most directly applicable to patients who receive first-line chemotherapy due to aggressive disease features or limited access to CDK4/6 inhibitors, and extrapolation to CDK4/6-pretreated populations should be made with caution. Future studies are warranted to evaluate the role of fulvestrant-based maintenance strategies following prior CDK4/6 inhibitor exposure. Third, the study population was relatively endocrine-sensitive, which may have favored endocrine-based maintenance and amplified the benefit observed with fulvestrant. Therefore, the findings are most applicable to this subgroup, and extrapolation to less endocrine-responsive populations should be made with caution. Another important limitation of the present report is the absence of mature quality-of-life (QoL) and patient-reported outcome (PRO) data. Additionally, the definition of “aggressive disease” used in this study is not fully standardized and is somewhat subjective. Prior capecitabine exposure in a subset of patients may have enriched for capecitabine-sensitive disease. Although balanced between arms, residual confounding due to intrinsic chemosensitivity cannot be excluded. As such, findings derived from these subgroup analyses should be interpreted with caution and considered exploratory. Finally, future studies should explore the underlying mechanisms that contribute to the differential effectiveness of fulvestrant versus capecitabine, such as molecular biomarkers that may predict response. Despite these limitations, this study provides strong evidence that short-course chemotherapy followed by fulvestrant maintenance is a feasible and effective strategy for HR + /HER2- MBC. This approach may be particularly beneficial for patients in regions where CDK4/6 inhibitors are not readily available or affordable. In conclusion, the FAMILY trial provides robust evidence supporting the use of fulvestrant as an effective and well-tolerated maintenance therapy for patients with HR + /HER2- MBC who have achieved an objective response or disease control after first-line chemotherapy. This study highlights the potential of fulvestrant maintenance to improve outcomes while preserving quality of life, offering a valuable alternative to chemotherapy-based maintenance strategies. These findings offer significant implications for managing HR + /HER2- MBC, providing clinicians with an important option for long-term disease control, especially in settings where CDK4/6 inhibitors are inaccessible or cost-prohibitive. Further research is needed to confirm the long-term benefits of fulvestrant maintenance, explore its combination with other therapies (e.g., CDK4/6 inhibitors), and investigate biomarkers for individualized treatment strategies. Materials and methods Study design This was a prospective, open-label, multicenter, randomized controlled phase III trial conducted at 22 hospitals in mainland China to compare fulvestrant with capecitabine in HR + /HER2- patients who had achieved an objective response or disease control following first-line systematic treatment. All patients provided written informed consent before enrollment. The study was approved by the institutional review boards and ethics committees of all participating hospitals and was conducted in compliance with the Declaration of Helsinki (Supplementary Table S1). This trial is registered with ClinicalTrials.gov (NCT04263298) and the Chinese Clinical Trial Register (ChiCTR-IIR-17014036). Participants The main inclusion criteria were as follows: (1) female patients aged 18–75; (2) histologically confirmed MBC deemed unsuitable for surgical resection or radiotherapy with curative intent; (3) tumors confirmed to be HR-positive, HER2-negative tumors; (4) no evidence of disease progression, defined as complete response (CR), partial response (PR), or stable disease (SD), following 4-8 course of first-line chemotherapy regimen (capecitabine monotherapy was permitted as first-line chemotherapy, limited to 6 cycles); (5) patients with at least one measurable disease per Response Evaluation Criteria in Solid Tumors Version 1.1 (RECIST 1.1) or with nonmeasurable bone-only disease (i.e., blastic, lytic, or mixed); (6) an Eastern Cooperative Oncology Group (ECOG) performance status of 0-1 and a life expectancy of at least 3 months; and (7) adequate organ function, including adequate bone marrow function (WBC ≥ 3.0 × 109/L, absolute neutrophil count [ANC] ≥1.5 × 109/L, platelet ≥100 × 109/L), renal function (serum creatinine ≤1.5 times upper limit of normal (ULN) or creatinine clearance ER and PR positivity were defined per ASCO/CAP guidelines as ≥1% of tumor cell nuclei showing positive staining by immunohistochemistry.45 Notably, clinically aggressive disease was defined as symptomatic visceral metastases, rapid disease progression or impending visceral compromise, or severely symptomatic non-visceral disease, as determined by the investigator.10 Key exclusion criteria included (1) newly developed central nervous system (CNS) metastases or uncontrolled symptomatic CNS spread necessitating corticosteroids or antiepileptic drugs within 4 weeks prior to study enrollment; (2) a history of another active malignancy within the past 3 years; (3) prior ET for advanced or recurrent breast cancer; and (4) recurrent or metastatic disease during the first 2 years of adjuvant ET (defined as primary endocrine resistance). Detailed inclusion and exclusion criteria are provided in the study protocol, available in the Supplementary information. Random assignment Using a stratified envelope randomization method, eligible patients were randomized in a 1:1 ratio to receive either fulvestrant or capecitabine. Randomization was stratified based on the sensitivity to ET (secondary resistance vs. sensitive). ET-sensitive patients were defined as those who had de novo metastatic disease and received no prior ET or those who had a disease-free interval >12 months from the completion of neoadjuvant or adjuvant ET. Patients with secondary ET resistance were defined as those whose disease relapsed during adjuvant ET but after the first 2 years or relapsed within 12 months of completing adjuvant ET. The randomization sequence was computer-generated using SAS 9.2 software by an independent statistician and stored securely. To ensure allocation concealment, the sequence was placed into opaque, tamper-proof, and sequentially numbered envelopes, which were distributed to each site. During enrollment, investigators opened the next sequentially numbered envelope only after confirming patient eligibility, ensuring an unbiased allocation process. Treatment Patients in the fulvestrant group received fulvestrant at a dose of 500 mg, administered intramuscularly on days 0, 14, and 28, and every 28 days thereafter. Premenopausal patients could receive concurrent ovarian suppression, either with a gonadotropin-releasing hormone agonist or via surgical oophorectomy. Patients in the capecitabine group received capecitabine at a dose of 2000 mg/m2/day, divided into two doses taken orally in the morning and evening, continuously for 14 days, followed by a 7-day rest period, in a 21-day cycle. Maintenance treatment was continued until disease progression, death, the occurrence of unacceptable toxicity, or patient withdrawal of consent for any reason. Dose interruptions were allowed. Dose reductions were permitted for capecitabine as prespecified in the protocol but were not applicable for fulvestrant. Study end points The primary end point was investigator-assessed PFS, which was calculated from the time of random assignment to disease progression, death, or the latest follow-up. Key secondary end points included OS, tPFS, tOS, survival outcomes (including PFS, OS, tPFS and tOS) for patients with clinically aggressive disease, objective response rate (ORR), CBR, and safety and tolerability. ORR was calculated as the percentage of patients with measurable disease at baseline who achieved either CR or PR. CBR was defined as the proportion of patients who achieved a confirmed CR, PR, or SD lasting ≥24 weeks. OS was calculated from the time of random assignment to death or the latest follow-up for patients who were still alive. tPFS was defined as the interval from the start of first-line chemotherapy to disease progression, death, or the most recent follow-up, whereas tOS was measured from the same starting point to death from any cause or the latest follow-up. Additionally, secondary end points, not reported in the current analysis, encompassed PROs and biomarker assessments; analyses of QoL measures using validated instruments (including the Functional Assessment of Cancer Therapy-Breast [FACT-B] and European Organization for Research and Treatment of Cancer Quality of Life Questionnaires-BR23 [QLQ-BR23] are ongoing and will be reported separately. Efficacy and safety assessments Tumor assessments were conducted using imaging techniques (CT or MRI) at baseline and subsequently every 8 weeks, with a ± 7-day window. Treatment response was categorized as CR, PR, SD, or progressive disease (PD) based on RECIST 1.1. Safety and tolerability assessments included AEs and serious AEs (SAEs) throughout the study. AEs were graded in accordance with the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03 (CTCAE 4.03). Laboratory parameters, electrocardiograms, physical examinations, and vital signs were systematically measured at prespecified intervals throughout the study. Follow-up assessments were conducted at three-month intervals for 18 months after the last cycle of study treatment. For patients who discontinued treatment early for reasons other than disease progression, efficacy assessments were performed every 6 weeks until 24 weeks after study initiation or until disease progression. Subsequently, follow-up visits were scheduled every 9 weeks until disease progression, death, or the end of the study. Statistical analysis The study hypothesized that the median PFS for the capecitabine maintenance therapy group would be 9.9 months, based on a reported first-line PFS of 14.8 months,46 with the expected duration of initial chemotherapy (4.9 months) subtracted to reflect the post-chemotherapy maintenance period. It was further hypothesized that fulvestrant maintenance would reduce the risk of disease progression by 35% (hazard ratio=0.65), corresponding to an expected median PFS of 15.2 months. A 1:1 randomization ratio was used, with a one-sided significance level (α) of 0.025 and a statistical power (1-β) of 80%. It was estimated that patient enrollment would be completed within 60 months, and the last enrolled participant would be followed for at least 24 months to observe PFS. Based on these parameters, 94 patients per group were needed, as calculated using the Power and Sample Size Program. Considering a 10% dropout or exclusion rate, each group will require 105 subjects, for a total of 210 subjects. The final analysis was planned after observing 173 PFS events (90 events in the control group and 83 events in the experimental group). Efficacy outcomes were analyzed based on the FAS, comprising all randomized participants who received at least one dose of the study drug. Safety assessment was conducted in the SS, including all participants with at least one dose and evaluable safety data. The Kaplan‒Meier method was used to generate survival estimates for PFS and OS, while the log-rank test assessed between-group differences. Hazard ratios and 95% confidence intervals (CIs) were derived from a stratified Cox proportional hazards model. Statistical significance was concluded if the upper bound of the 95% CI for the hazard ratio of the study treatment was less than 1 for the primary endpoint PFS. Subgroup analyses for PFS were performed using Cox regression to calculate the hazard ratio (95% CI), and the results are presented in forest plots. The duration of follow-up was calculated with reversed Kaplan‒Meier methods. The primary endpoint, PFS, was analyzed with a one-sided significance level (α = 0.025). All other statistical tests were conducted at a two-sided significance level (α = 0.05). All analyses were conducted using R version 4.2.2. The full statistical analysis plan is available with the protocol. Supplementary information CONSORT checklist Supplementary Material Supplementary Appendix-Protocol CONSORT checklist Supplementary Material Supplementary Appendix-Protocol