Main

Breast cancer (BC) is the most frequently diagnosed cancer in women around the world, including in South Korea and Taiwan, and is the second most commonly diagnosed cancer in women in China1,2,3,4. The most common type of BC is hormone receptor-positive human epidermal growth factor receptor 2-negative (HR+HER2−), which accounts for approximately 70% of BC cases5,6. Current standard-of-care therapy for HR+HER2−) metastatic BC (mBC) is endocrine therapy combined with a cyclin-dependent kinase 4/6 inhibitor (CDK4/6i) as first line, followed by sequential lines of single-agent chemotherapy once endocrine resistance develops7,8,9,10. Single-agent chemotherapy is associated with limited efficacy and high toxicity, highlighting the unmet need for additional treatment options in patients with HR+HER2− mBC11,12,13,14,15.

Sacituzumab govitecan (SG) is a trophoblast cell surface antigen 2 (Trop-2)-directed antibody–drug conjugate. SG selectively targets cells expressing Trop-2, which is commonly expressed in BC16. It is then internalized and delivers SN-38, an active metabolite of irinotecan, via a hydrolysable linker17,18,19. SG is approved in the United States and in multiple other countries for the treatment of HR+HER2− mBC following endocrine therapy and at least two additional systemic chemotherapies in the metastatic setting based on the global phase 3 TROPiCS-02 study20,21. SG is also approved in multiple countries for the treatment of triple-negative mBC for patients who have received at least two prior systemic therapies, including at least one in the metastatic setting20,21. SG has also received accelerated approval in the United States for the treatment of metastatic urothelial cancer following chemotherapy and immunotherapy20.

In the TROPiCS-02 study (NCT03901339), SG was compared with chemotherapy treatment of physician’s choice. At the primary progression-free survival (PFS) analysis (data cutoff date 3 January 2022), SG demonstrated statistically significant improvement in the primary endpoint of PFS versus chemotherapy, with a 34% reduction in risk of disease progression or death (hazard ratio (HR) of 0.66, 95% confidence interval (CI) 0.53–0.83; P = 0.0003) and median of 5.5 versus 4.0 months22. At a subsequent prespecified interim analysis (data cutoff date 1 July 2022), SG demonstrated statistically significant and clinically meaningful improvement in overall survival (OS), with a median of 14.4 versus 11.2 months (HR of 0.79, 95% CI 0.65 to 0.96; P = 0.020), for SG versus chemotherapy23. Treatment with SG also resulted in higher objective response rate (ORR) and median duration of response (DoR) versus chemotherapy, which were 21% versus 14% and 8.1 versus 5.6 months, respectively23. The safety profile of SG was manageable and similar to earlier clinical trials in BC and other tumors22,23,24,25,26. The patients treated with SG experienced grade ≥3 treatment-emergent adverse events (TEAEs) at higher rates than with chemotherapy (74% versus 60%), but the rates of TEAEs leading to treatment discontinuation were similar (6% versus 4%)23.

TROPiCS-02 primarily enrolled non-Asian patients, with Asian patients making up 3% of the study population22. Thus, the benefit–risk profile for SG in Asian patients with HR+HER2− mBC has not been adequately characterized. The phase 3, randomized EVER-132-002 study (NCT04639986) was initiated to determine the efficacy and safety of SG in Asian patients with endocrine-resistant, pretreated HR+HER2− mBC.

Results

Patient characteristics

Of the 485 patients screened between 9 December 2020 and 2 December 2022, 331 were randomized at 41 study sites in China, Republic of Korea and Taiwan to receive SG (n = 166) or chemotherapy (n = 165; eribulin, n = 131 (79%); vinorelbine, n = 13 (8%); capecitabine, n = 11 (7%); gemcitabine, n = 10 (6%); Fig. 1). Two patients were randomized to the chemotherapy group but did not receive any study drug. One patient was randomized to the SG group but was erroneously administered chemotherapy; this patient was included in the SG group for efficacy outcomes and the chemotherapy group for safety outcomes. At the data cutoff date of 30 April 2023, 21 patients were continuing study drug, 6% (n = 10) in the SG group and 7% (n = 11) in the chemotherapy group.

Fig. 1: CONSORT diagram.
figure 1

Of 485 patients screened, 154 were not randomized to a treatment group. The remaining 331 patients were randomized to the SG or chemotherapy groups. aChemotherapy selected at randomization: eribulin (n = 131), vinorelbine (n = 13), capecitabine (n = 11) or gemcitabine (n = 10). bOne patient was randomized to the SG group but was erroneously administered chemotherapy. cOther reasons included clinical deterioration, pregnancy or administrative problems. dOther reasons included clinical deterioration, pregnancy, administrative problems or patient nonadherence/protocol violations.

Full size image

Patient demographics and baseline characteristics were generally well balanced across treatment groups. The median age was 52 years. Most patients had visceral metastases at baseline (88% SG, 89% chemotherapy). Approximately half of patients had received prior CDK4/6i therapy in the metastatic setting (49% SG, 48% chemotherapy). In both treatment groups, 56% of patients had received two prior chemotherapy regimens for treatment of metastatic disease, and 44% had received three to four prior regimens (Table 1).

Table 1 Patient demographics and baseline characteristics
Full size table

Most patients had received prior endocrine therapy in the metastatic setting (92% SG, 90% chemotherapy), and 37% of patients in the SG group and 41% in the chemotherapy group had previously received targeted therapy in the metastatic setting (Extended Data Table 1).

Efficacy

Median follow-up duration was 13.4 months (range 0.1–28.7) in all randomized patients, 14.5 months (range 1.3–27.8) in the SG group and 12.7 months (range 0.1–28.7) in the chemotherapy group. The primary endpoint of PFS per blinded independent central review (BICR) was met, with a 33% reduction in the risk of disease progression or death (HR of 0.67, 95% CI 0.52 to 0.87; P = 0.0028) (Fig. 2a). The median PFS per BICR was 4.3 months (95% CI 4.1 to 5.7) in the SG group versus 4.2 months (95% CI 2.8 to 4.2) in the chemotherapy group. The PFS rate at 6 months was 41% versus 24% and at 12 months was 17% versus 9% with SG versus chemotherapy, respectively. The results for PFS per investigator (INV) were generally consistent with the respective results per BICR (HR of 0.61, 95% CI 0.48 to 0.79; P = 0.0001), with a longer median PFS demonstrated in the SG group (5.7 months (95% CI 4.3 to 7.9)) versus the chemotherapy group (4.2 months (95% CI 2.9 to 4.3)) (Fig. 2b). PFS benefit was generally consistent across multiple prespecified subgroups, including by age, baseline Eastern Cooperative Oncology Group performance status, HER2 status, presence of visceral metastasis, number of prior chemotherapy regimens for treatment of metastatic disease and prior CDK4/6i treatment (Fig. 3).

Fig. 2: Kaplan–Meier estimates of PFS and OS.
figure 2

a, Kaplan–Meier curve of PFS per BICR in patients treated with SG (n = 166) versus chemotherapy (n = 165). b, Kaplan–Meier curve of PFS per INV in patients treated with SG versus chemotherapy. c, Kaplan–Meier curve of OS in patients treated with SG versus chemotherapy. The median PFS and OS were calculated using the Kaplan–Meier method. The HR values were estimated using a stratified Cox proportional-hazards model, and the P values were calculated using a stratified log-rank test. Dashed blue and grey lines indicate median values, and dashed black lines indicate landmark time points. aStratified log-rank P value. NE, not estimable.

Full size image
Fig. 3: Subgroup analysis of PFS.
figure 3

PFS per BICR with SG versus chemotherapy was analyzed across multiple subgroups. The median PFS was calculated using the Kaplan–Meier method, and HR values were estimated using a Cox proportional-hazards model. aHER2-low was defined as IHC score of 1+ or score of 2+ with a negative fluorescence in situ hybridization result; HER2 IHC0 was defined as an IHC score of 0.

Full size image

A 36% reduction in the risk of death was observed with SG versus chemotherapy (HR of 0.64, 95% CI 0.47 to 0.88; P = 0.0061). The median OS with SG versus chemotherapy was 21.0 (95% CI 16.5 to not estimable) versus 15.3 months (95% CI 13.2 to 18.4). The OS rate for SG versus chemotherapy was 76% versus 62% at 12 months and 41% versus 30% at 24 months (Fig. 2c). The ORR per BICR was 20% with SG and 15% with chemotherapy. Median DoR per BICR was 5.3 months (SG) versus 5.2 months (chemotherapy). Responses per INV assessment were consistent with responses per BICR (Table 2). SG generally provided consistent OS benefit versus chemotherapy across prespecified subgroups (Extended Data Fig. 1).

Table 2 Objective response summary
Full size table

When patients were divided by Trop-2 H-score (<100 versus ≥100), the HR for PFS per BICR with SG versus chemotherapy was 0.91 (95% CI 0.65 to 1.28) in the Trop-2 <100 subgroup and was 0.44 (95% CI 0.29 to 0.68) in the ≥100 subgroup. The HR for OS with SG versus chemotherapy was 0.72 (95% CI 0.47 to 1.10) for Trop-2 < 100 and 0.61 (95% CI 0.37 to 1.01) for ≥100 (Fig. 3 and Extended Data Fig. 1).

Safety

Among patients in the safety population, the median duration of treatment with SG was 5.1 months (range 0.03–24.9) and with chemotherapy was 3.3 months (range 0.03–28.1). The patients were administered a median of seven cycles of SG compared with five cycles with chemotherapy, and the median relative dose intensity for SG was 100% (Extended Data Table 2).

All patients in both treatment groups experienced any-grade TEAEs. Grade ≥3 TEAEs in the SG versus chemotherapy groups were reported in 82% versus 70% of patients, respectively. The most common TEAEs of any-grade with SG versus chemotherapy were neutropenia (88% versus 78%), anemia (71% versus 55%) and leukopenia (68% versus 63%). The most common grade ≥3 TEAEs with SG versus chemotherapy were neutropenia (69% versus 62%), leukopenia (42% versus 37%) and anemia (18% versus 6%) (Table 3). Febrile neutropenia (any grade and grade ≥3) was observed in 5% of patients in the SG group and 4% in the chemotherapy group. TEAEs leading to dose interruption with SG versus chemotherapy occurred at a rate of 68% versus 40%, and TEAEs leading to treatment discontinuation 3% versus 4%. TEAEs leading to dose reduction were reported in 25% of patients in the SG group and 32% in the chemotherapy group, and treatment-emergent serious AEs occurred in 23% and 20% of patients, respectively. There was one TEAE leading to death observed in both treatment groups, one instance of septic shock (SG) and one of asphyxia (chemotherapy). Neither of these events were assessed as treatment-related by the INV. The patient with septic shock had been receiving central parenteral nutrition due to a bowel obstruction starting approximately 4 months before initiation of SG, and central parenteral nutrition-related infection was considered the reason for septic shock by the INV.

Table 3 Safety summary
Full size table

Exposure-adjusted incidence rate (EAIR) was similar between the SG and chemotherapy groups for any-grade and grade ≥3 TEAEs. EAIR was also similar between the treatment groups for serious TEAEs, TEAEs leading to treatment discontinuation and TEAEs leading to death. EAIR for TEAEs leading to dose interruption was higher with SG than chemotherapy (EAIR difference versus chemotherapy 1.18; 95% CI 0.51 to 1.86). EAIR for TEAEs leading to dose reduction was lower with SG versus chemotherapy (EAIR difference versus chemotherapy −0.58; 95% CI −1.00 to −0.20; Extended Data Table 3).

Polymorphisms in uridine diphosphate glucuronosyltransferase 1A1 (UGT1A1) have been associated with an increased incidence of specific AEs during treatment with some systematic anticancer agents27,28,29,30. In this study, the most common UGT1A1 genotypes in patients receiving SG were *1/*1 (wild type; n = 77), *1/*6 (n = 40) and *1/*28 (n = 18). Patients in the *1/*1 subgroup experienced numerically lower rates of grade ≥3 TEAEs, TEAEs leading to treatment interruption and TEAEs leading to dose reduction when compared with *1/*6 and *1/*28. Patients with heterozygous genotypes (*1/*6, *1/*28) had increased rates of any-grade neutropenia, anemia and diarrhea compared with *1/*1, and grade ≥3 neutropenia and diarrhea were also more common in heterozygous patients (Extended Data Table 4). Rates of febrile neutropenia were similar for *1/*1 versus *1/*6 versus *1/*28 (4% versus 5% versus 6%, respectively, for any-grade and grade ≥3 AEs).

QoL

Most patients (96%, n = 318) were evaluable for quality of life (QoL) assessments using European Organisation for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ-C30 v3.0) and European Quality of Life (EuroQOL) five-dimension five-level (EQ-5D-5L) questionnaire. The median time to deterioration (TTD) was longer with SG than with chemotherapy across the EORTC QLQ-C30 global health status/QoL, pain, fatigue and physical functioning measures (Extended Data Table 5). SG demonstrated a significant extension in TTD over chemotherapy on the physical functioning and pain measures when death was censored. However, when considering death as an event, a significant extension in TTD was observed only in the physical functioning measure; for the other three measures, the TTD was longer for SG, though the differences were not significant between SG and chemotherapy. In EQ-5D-5L patient responses, mobility, self-care, usual activities, pain/discomfort and anxiety/depression scores decreased for patients in both treatment groups from baseline to end of treatment. The percentage of patients with severe problems was generally numerically higher in the chemotherapy group versus the SG group at end of treatment, and the percentage of patients with no problems was higher in the SG group versus the chemotherapy group (Extended Data Fig. 2).

Discussion

Due to the limited options available and poor outcomes for patients with previously treated HR+HER2− mBC, an unmet need exists for treatments that improve outcomes in this patient population. SG demonstrated statistically significant improvement in PFS and clinically meaningful improvement in OS compared with chemotherapy in Asian patients with endocrine-resistant, chemotherapy-pretreated HR+HER2− mBC in the phase 3 EVER-132-002 trial. The primary endpoint of PFS per BICR was met, with a 33% reduction in risk of disease progression or death (HR of 0.67, P = 0.0028). The median PFS values were similar between SG (4.3 months) and chemotherapy (4.2 months) due to a convergence of the curves at the median time point and do not reflect the overall benefit that patients in the SG group received as measured by the HR. Importantly, nearly twice as many patients were alive and progression free in the SG group compared with the chemotherapy group at all prespecified landmark time points of 6, 9 and 12 months. PFS per INV review was generally consistent with PFS per BICR. SG also demonstrated a clinically meaningful improvement in OS versus chemotherapy, with a 36% reduction in the risk of death (HR of 0.64, P = 0.0061). Median OS was 5.7 months longer in the SG group (21.0 months) compared with the chemotherapy group (15.3 months). The Kaplan–Meier curves separated early and widened over time with a higher proportion of patients alive in the SG group at both landmark time points of 12 and 24 months. ORR and clinical benefit rate (CBR) per BICR were numerically higher in patients treated with SG, and BICR assessment was generally consistent with INV assessment results.

PFS and OS benefit with SG was observed across most predefined subgroups, including for each stratification factor (visceral metastases, prior treatment with CDK4/6i in the metastatic setting and number of prior chemotherapy regimens for treatment of metastatic disease). Notably, 49% of patients had received prior CDK4/6i in this study. This reflects the treatment landscape in China, South Korea and Taiwan at the time of trial design, in which first-line CDK4/6i treatment was less accessible for patients with HR+HER2− mBC. As current cancer care guidelines in Asia include first-line use of CDK4/6i for advanced HR+HER2− BC9,10, it is expected that the majority of patients in this population are likely to have received prior CDK4/6i treatment. The patients in EVER-132-002 experienced clinical benefit regardless of whether prior treatment with a CDK4/6i was received.

Safety in this trial was consistent with the safety profile observed in previous global studies of SG23,24,26, and TEAEs were generally manageable. No new safety signals were identified. Grade ≥3 TEAEs were higher in the SG group compared with the chemotherapy group. The incidence of TEAEs leading to dose reduction was lower in the SG versus chemotherapy group, and the incidence of TEAEs leading to discontinuation and death was similar in the SG and chemotherapy groups.

TTD was numerically improved with SG versus chemotherapy for global health status/QoL, pain, fatigue and physical functioning scales, and this improvement was significant for physical functioning and for pain when death was censored. EQ-5D-5L patient responses showed decreases in all dimensions analyzed from baseline to end of treatment in both treatment groups; however, patients in the SG group generally had numerically higher rates of no problems and lower rates of severe problems compared with the chemotherapy group.

The results of EVER-132-002 were consistent with the results of TROPiCS-02, a global phase 3 study that evaluated SG as a treatment for HR+HER2− mBC and was conducted primarily in non-Asian patients. The eligibility criteria were generally comparable between these studies, with similar median age and a majority of patients with visceral metastases at baseline. One exception to this was the percentage of patients who had received previous CDK4/6i treatment. In TROPiCS-02, 99% of patients had received prior CDK4/6i, compared with 49% of patients in EVER-132-002, which reflected the local treatment paradigm for HR+HER2− mBC at the time of the study enrollment. In an indirect comparison, PFS benefit of SG versus chemotherapy between EVER-132-002 and TROPiCS-02 was comparable with HR 0.67 and HR 0.66, respectively22. OS benefit from SG versus chemotherapy was also comparable between the studies, with HRs of 0.64 and 0.79, respectively23. Patients in EVER-132-002 experienced clinical benefit regardless of whether prior treatment with a CDK4/6i was received. Efficacy data from patients who received prior CDK4/6i in EVER-132-002 were also consistent with those in the overall study population of TROPiCS-02, all of whom had received prior CDK4/6i treatment. Overall, efficacy results from the intent-to-treat population were consistent between EVER-132-002 and TROPiCS-02. Additional OS follow-up is ongoing in the EVER-132-002 study.

Subgroup analyses of EVER-132-002 were also generally consistent with previous results from TROPiCS-02. While SG is targeted to Trop-2, previous studies have indicated that SG provided clinical benefit versus chemotherapy in patients with previously treated HR+HER2− mBC and low Trop-2 expression. In a post hoc analysis of TROPiCS-02, PFS and OS benefit were observed with SG versus chemotherapy in both the Trop-2 <100 and ≥100 subgroups31,32. In EVER-132-002, OS benefit was observed with SG versus chemotherapy in both Trop-2 subgroups. PFS HR values numerically favored SG over chemotherapy in both Trop-2 expression subgroups, indicating that SG provided benefit even in patients with low Trop-2 expression. The sample size in these subgroups was small and the study was not powered to detect differences between subgroups, which limits the interpretation of this analysis.

Differences in rates of TEAEs by UGT1A1 genotype were observed in patients with previously treated HR+HER2− mBC in an exploratory analysis of TROPiCS-02. Patients with the *28/*28 genotype for UGT1A1 experienced numerically higher rates of grade ≥3 TEAEs, TEAEs leading to discontinuation, any-grade anemia and diarrhea or neutropenia of grade ≥3, when compared with patients with *1/*1 (wild type) or *1/*28 genotypes33. The most common UGT1A1 genotypes differed between the studies, with *1/*1, *1/*6 and *1/*28 being the most common in EVER-132-002, while *1/*1, *1/*28 and *28/*28 were the most common in TROPiCS-02 (ref. 33). Patients with heterozygous genotypes consistently experienced higher rates of grade ≥3 TEAEs, TEAEs leading to dose reduction and grade ≥3 neutropenia and diarrhea than those with *1/*1 (wild type) genotypes. However, the number of patients with heterozygous genotypes was relatively small, which may limit interpretation of these results.

The results of HRQoL analyses from EVER-132-002 generally aligned with those from TROPiCS-02. In both trials, median TTD was numerically higher with SG than chemotherapy across all categories analyzed. This difference was significant in global health status/HRQoL and fatigue in TROPiCS-02, while in EVER-132-002, it was significant in pain and physical functioning23. These differences may be explained due to differences in the patient population between the two trials.

One potential limitation of this study is the open-label design. Despite this, only two patients who were assigned to the chemotherapy group withdrew from the study before receiving any study treatment.

The patients in EVER-132-002 are representative of a real-world Asian patient population. This is the first clinical trial of SG to show substantial improvement in both PFS and OS in this patient population. The efficacy benefit of SG versus chemotherapy and the manageable safety profile in the pivotal EVER-132-002 study are consistent with findings of the global TROPiCS-02 study and support the use of SG as a new treatment option for Asian patients with endocrine-resistant, previously treated HR+HER2− mBC.

Methods

Ethics statement

EVER-132-002 was compliant with Declaration of Helsinki and International Council for Harmonisation Good Clinical Practice guidelines and was approved by national regulatory authorities, as well as each investigational sites’ ethics committee or review board. The investigational sites that approved the study protocol were Cancer Hospital Chinese Academy of Medical Science; Chinese PLA General Hospital; Peking University People’s Hospital; Jilin Cancer Hospital; The First Hospital of Jilin University; Chongqing University Cancer Hospital; West China Hospital, Sichuan University; Fujian Medical University Union Hospital; Guangdong Provincial People’s Hospital; Sun Yat Sen Memorial Hospital of Sun Yat Sen University; Sun Yat Sen University Cancer Center; Sir Run Run Shaw Hospital Zhejiang University School of Medicine; Zhejiang Cancer Hospital; Anhui Provincial Hospital; The Second Hospital of Anhui Medical University; Shandong Cancer Hospital; Yunnan Cancer Hospital; Linyi Cancer Hospital; Jiangsu Province Hospital; Nanjing Drum Tower Hospital; Shanghai General Hospital; Tianjin Medical University Cancer Institute and Hospital; Hubei Cancer Hospital; Union Hospital, Tongji Medical College, Huazhong University of Science and Technology; The First Affiliated Hospital of Xi’an Jiaotong University; Henan Cancer Hospital; Cancer Hospital of Xinjiang Medical University; Dong-A University Hospital; Seoul National University Bundang Hospital; Asan Medical Center; Korea University Anam Hospital; Samsung Medical Center; Seoul National University Hospital; Severance Hospital of Yonsei University Health System; Ajou University Hospital; Changhua Christian Medical Foundation Changhua Christian Hospital; Kaohsiung Medical University Chung-Ho Memorial Hospital; China Medical University Hospital; National Cheng Kung University Hospital; National Taiwan University Hospital; Taipei Veterans General Hospital; Tri-Service General Hospital; and Chang Gung Memorial Hospital, Linkou. All patients gave written informed consent.

Patients

Patients from mainland China, Republic of Korea and Taiwan were eligible for enrollment if they had histologically or cytologically confirmed HR+HER2− metastatic or locally recurrent inoperable BC, with most recently available or newly obtained tumor biopsy from a locally recurrent or metastatic site. Patient sex was self-reported and recorded by the study site in the study database. Archival or newly acquired tumor tissue in a formalin-fixed, paraffin embedded block or adequate unstained slides from a metastatic or recurrent site was required. HR+ was defined as ≥1% of cells expressing hormonal receptors (estrogen and/or progesterone) by immunohistochemistry (IHC), and HER2− was defined as IHC0, IHC1+ or IHC2+ and in situ hybridization negative. Patients were required to have received two to four prior systemic chemotherapy regimens for mBC and to be eligible for one of the options for chemotherapy of INV’s choice. Adjuvant or neoadjuvant chemotherapy was counted as one of the required prior regimens if development of unresectable, locally advanced or mBC occurred within a 12 month period after completion of chemotherapy. Previous treatment included at least one taxane and at least one anticancer endocrine therapy in any setting; prior use of CDK4/6i was not mandatory. Documented measurable disease by Response Evaluation Criteria in Solid Tumors v1.1 (by computed tomography or magnetic resonance imagery) following the most recent anticancer therapy was required. The patients were required to be aged ≥18 years and to provide written informed consent, Eastern Cooperative Oncology Group performance status of 0 or 1 and to be recovered from all prior treatment-related toxicities to grade ≤1 per the Common Terminology Criteria for Adverse Events v5.0 (except alopecia or stable sensory neuropathy, which could be grade ≤2).

Exclusion criteria included previous treatment with topoisomerase 1 inhibitors, known brain metastases, second active malignancy within 3 years before providing informed consent (nonmelanoma skin cancer and histologically confirmed completed excision of carcinoma in situ excepted), human immunodeficiency virus positivity, active hepatitis B or C virus infection, known history of unstable angina/myocardial infarction/congestive heart disease within 6 months before first dose, clinically significant cardiac arrythmia requiring antiarrythmia therapy or left ventricular ejection fraction <50%, known history of clinically significant active chronic obstructive pulmonary disease or other moderate-to-severe chronic respiratory illness within 6 months before first dose, active chronic inflammatory bowel disease (ulcerative colitis or Crohn’s disease)/clinically significant gastrointestinal bleeding/intestinal obstruction/gastrointestinal perforation within 6 months before first dose, active serious infection requiring systemic antibiotic use within 7 days before first dose, high-dose systemic corticosteroid use within 2 weeks before first dose, scheduled surgery that would delay study treatment during study, receipt of live vaccine within 30 days before first dose, other concurrent medical or psychiatric conditions that may confound study interpretation or prevent completion of study procedures or follow-up examinations, known hypersensitivity or intolerance to study treatments or any of the excipients, anticancer treatment with chemotherapy/radiation/small molecule targeted therapy/endocrine therapy within 2 weeks before first dose, biological therapy within 4 weeks before first dose, current treatment in another clinical study or use of any investigational drug or device within five half-lives or 4 weeks before first dose (whichever is longer), pregnancy or lactation and women of childbearing potential/fertile men unwilling to use highly effective contraception during the study and up to 6 months after treatment discontinuation for women of childbearing potential and 3 months for fertile men.

Study design

Patients were randomly assigned at 1:1 to receive SG (10 mg kg−1 intravenously days 1 and 8 of every 3 week cycle) or chemotherapy treatment of physician’s choice (eribulin, capecitabine, gemcitabine or vinorelbine, with dosing and schedule per package insert depending on region). Randomization was stratified according to presence of visceral metastases (yes versus no), number of prior chemotherapy regimens for the treatment of metastatic disease (two versus three/four) and prior CDK4/6i treatment in the metastatic setting (yes versus no). Patients were randomized via an interactive web-based response system. The randomization list was generated by a designated vendor using stratified blocked randomization with a block size of four.

Endpoints

The primary endpoint was PFS per BICR using Response Evaluation Criteria in Solid Tumors v1.1. PFS was defined as time from randomization to progressive disease (PD) or death, whichever occurred earlier.

Secondary endpoints included OS, ORR per BICR, DoR per BICR, CBR per BICR, safety and QoL. Efficacy by INV assessment was not a prespecified endpoint but was a planned supportive analysis per the protocol. OS was defined as time from randomization to death from any cause. ORR was the sum of complete response (CR) and partial response (PR), while CBR was the sum of CR, PR and stable disease (SD) with duration ≥6 months. DoR was defined as time from first tumor response (CR or PR) to PD or death, whichever was first. The exploratory endpoints included Trop-2 expression level.

Assessments

Stratification factor information was gathered from an interactive web-based response system. The tumor measurements by computed tomography or magnetic resonance imagery were performed every 6 weeks up to 54 weeks. After 54 weeks, measurements were performed every 12 weeks. Safety was assessed according to National Cancer Institute Common Terminology Criteria for Adverse Events v5.0, and the events were coded using MedDRA v26.0. QoL was assessed using EORTC QLQ-C30 v3.0 and the EuroQOL EQ-5D-5L questionnaire. Blood and tumor samples for biomarker analyses were collected before the first dose of SG. The UGT1A1 genotype was tested through gene sequencing of whole blood, and tumor Trop-2 expression was evaluated using immunohistology staining of fresh or archived specimens. Membrane Trop-2 expression was determined using a validated research IHC assay at a central laboratory. The data were categorized using the H-score, which is the sum of percent staining weighted by staining intensity.

Statistical analysis

Sample size was calculated on the basis of the primary endpoint of PFS per BICR. Assuming an HR of 0.70 (median 5.3 versus 3.7 months), approximately 250 PFS events were needed to detect a statistically significant difference between the two groups at a two-sided 0.05 significance level and 80% power. Based on an estimated 12 month enrollment and 24 month total study duration, it was estimated that approximately 330 patients would need to be randomized.

There was no interim analysis planned for the primary endpoint. The primary analysis of PFS was conducted on the basis of a total of 244 PFS events observed as of the data cutoff date (30 April 2023). At the time of the primary analysis of PFS, all secondary endpoints were analyzed and reported. Only the primary endpoint of PFS was formally tested for the study.

All statistical analyses were performed using SAS (SAS Institute, version 9.4 or later, SAS Institute). Comparisons of PFS and OS between the treatment groups were performed using a stratified log-rank test (with the three stratification factors used during randomization). Median PFS and OS were determined using the Kaplan–Meier method, and their associated 95% CIs were calculated by the Brookmeyer and Crowley method with log–log transformation. The Kaplan–Meier estimates of PFS and OS were plotted over time. HRs were estimated using a Cox proportional-hazards model stratified by the stratification factors used in randomization. ORR and CBR were calculated with exact 95% CIs using the Clopper–Pearson method, and a comparison between treatment groups was performed with a stratified Cochran–Mantel–Haenszel test. Kaplan–Meier analyses were performed for DoR.

Safety was summarized by treatment group using descriptive statistics. TEAEs were defined as any AE that started on or after the first dose date and up to 30 days after the last dose date. EAIR was defined as number of patients who had at least one event divided by total patient-years of exposure from first dose date to first onset of event (for those who had an event) or to data cutoff date (for those who did not have an event and were continuing study drug) or last dose (for those who did not have an event and discontinued study drug).

Results from the EQ-5D-5L questionnaire were descriptively evaluated. TTD was measured using EORTC QLQ-C30 and was defined as time from randomization to first date patient reached ≥10-point deterioration from baseline or death, whichever occurred earlier. TTD was also analyzed considering death as censored. HRs and 95% CIs were calculated for TTD using a stratified Cox proportional-hazards model.

The intent-to-treat population consisted of all patients who were randomized to a treatment group. All patients who received at least one dose of study treatment were included in the safety analysis population. The HRQoL-evaluable and the EQ-5D-5L-evaluable populations were defined as patients in the intent-to-treat population with an evaluable assessment at baseline and at least one evaluable assessment at postbaseline visits.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.