Eligibility for adjuvant CDK4/6 inhibitors in HR+/HER2- early breast cancer: insights from a large cohort in central south China.

Introduction
Breast cancer (BC) is the most commonly diagnosed malignancy in women worldwide, with hormone receptor-positive (HR+) disease accounting for approximately 70% of all cases (1). Despite significant advances in endocrine therapy (ET), the risk of recurrence remains a major challenge, particularly for patients with high-risk clinicopathological features. The introduction of cyclin-dependent kinase 4/6 (CDK4/6) inhibitors has revolutionized the adjuvant treatment landscape, offering promising reductions in recurrence rates for HR+ early breast cancer (EBC) (2). Two landmark trials, MonarchE (abemaciclib) and NATALEE (ribociclib), have established the efficacy of CDK4/6 inhibitors in combination with ET, demonstrating their ability to reduce relapse risk in distinct high-risk patient populations (3).
The MonarchE trial enrolled patients with high-risk clinicopathological features, defined as ≥4 positive axillary lymph nodes (LNs) or 1–3 positive LNs accompanied by either tumor size ≥5 cm or histologic grade 3 (4). In contrast, the NATALEE trial employed broader inclusion criteria, recruiting stage IIB–III patients as well as selected high-risk stage IIA subgroups regardless of nodal status (5). These differing criteria reflect an evolving understanding of “high risk” in EBC, integrating tumor biology markers such as Ki-67 and histologic grade alongside traditional anatomical staging. However, applying these trial-based criteria to heterogeneous real-world populations remains challenging. Most existing studies have focused on validating trial outcomes within controlled settings, with limited data characterizing real-world patients who fulfill these eligibility criteria (6,7). Moreover, the degree of overlap between the NATALEE and MonarchE eligibility populations—and the resulting implications for clinical treatment decisions—has not been fully elucidated. This knowledge gap is significant given the financial costs, logistical complexities, and potential toxicities associated with CDK4/6 inhibitor therapy, underscoring the need for precise patient stratification (8,9). Currently, data on the proportion of patients meeting these trial criteria in real-world clinical practice—especially in economically moderate regions of China are scarce. Additionally, patients in these settings often present with a younger age at diagnosis and distinct molecular profiles compared to Western cohorts, which may influence both eligibility and therapeutic response.
Therefore, we analyzed the prevalence and clinicopathological characteristics of HR+ EBC patients eligible for the NATALEE and MonarchE trials within a large Chinese cohort. This study provides valuable insights into optimizing the use of CDK4/6 inhibitors in the Chinese population, bridging the gap between trial-defined criteria and real-world clinical practice. We present this article in accordance with the STROBE reporting checklist (available at https://tbcr.amegroups.com/article/view/10.21037/tbcr-25-27/rc).

Methods

Study design and population
This retrospective study included patients diagnosed with HR+ BC who were treated at the Department of Breast Surgery, Xiangya Hospital, Central South University, between January 2015 and December 2021. Patients were identified through the hospital’s electronic medical records (EMRs) and pathology databases. Eligibility criteria required histologically confirmed invasive HR+ BC, defined as estrogen receptor (ER) and/or progesterone receptor (PR) expression ≥1% by immunohistochemistry (IHC), with complete staging data according to the 8th edition of the American Joint Committee on Cancer (AJCC) guidelines (10). Exclusion criteria comprised: (I) non-HR+ subtypes, including HER2-positive and triple-negative tumors; (II) ductal carcinoma in situ (DCIS) without invasive components; and (III) incomplete clinicopathological or follow-up data.

Data collection and variables
Demographic and clinicopathological variables were extracted from EMRs, including age at diagnosis, tumor size (T stage), nodal status (N stage), histologic grade (using the modified Bloom-Richardson system), Ki-67 proliferation index (with a cutoff of ≥20% indicating high proliferation), and molecular subtype classification. To ensure accuracy and consistency with AJCC 8th edition staging, two independent breast pathologists re-evaluated tumor staging, resolving discrepancies by consensus.

Eligibility criteria for MonarchE and NATALEE trials
Patients were stratified into cohorts based on trial-specific inclusion criteria (11). For the MonarchE trial, eligibility included: Cohort 1—patients with ≥4 positive axillary LNs (pN2-3) or 1–3 positive nodes (pN1) combined with either tumor size ≥5 cm (T3) or histologic grade 3 (G3); Cohort 2—patients with 1–3 positive nodes (pN1), centrally tested Ki-67 ≥20%, tumor size <5 cm, and histologic grade <3. Patients meeting criteria for either cohort were classified as MonarchE-eligible. NATALEE trial eligibility comprised stage III (any T, N1–3), stage IIB (T3N0 or T2N1), or stage IIA (T2N0) with high-risk features including histologic grade 3 (G3), Ki-67 ≥20%, or high genetic risk (Oncotype DX Recurrence Score ≥26; Prosigna/PAM50 high risk; MammaPrint or EndoPredict EPclin high risk scores). For stage IIA (T2N0) patients, eligibility required the presence of either G3, Ki-67 ≥20%, or documented genetic risk. Due to the limited availability of genetic testing in this real-world cohort, genetic risk was inferred based on institutional guidelines, primarily through high-risk clinicopathological features. As a surrogate, Ki-67 was used to reflect tumor biology. We acknowledge that this substitution may introduce misclassification bias in patient risk stratification.

Subgroup analysis of N0 patients in NATALEE
Within the NATALEE-eligible cohort, node-negative (N0) patients were further stratified according to tumor size and biological risk factors: T4N0, T3N0, or T2N0 tumors with histologic grade 2 (G2) and Ki-67 ≥20%, grade 3 (G3), or high genetic risk. Ki-67 index and histologic grading were validated through centralized testing when available.

Statistical analysis
Descriptive statistics summarized continuous variables as medians with ranges, and categorical variables as frequencies and percentages. As the normality assumption was not met, non-parametric methods were applied. Statistical analyses were performed using R software (version 4.4.1; R Foundation for Statistical Computing, Vienna, Austria). Categorical variables were compared between groups using Pearson’s Chi-squared test when all expected cell counts were ≥5. For contingency tables with sparse data or small expected frequencies, Fisher’s exact test with Monte Carlo simulation (100,000 replications) was applied to obtain accurate P values. Missing data, such as Ki-67 results from non-centralized tests, were addressed via complete-case analysis. Ki-67 data were available for the vast majority of patients. Only 13 patients (0.7%) in the overall cohort, 6 patients (0.5%) in the population meeting the NATALEE eligibility criteria, and 3 patients (0.6%) in the population meeting the MonarchE eligibility criteria had missing Ki-67 information. A complete-case analysis approach was applied, given the extremely low proportion of missing data. Flow charts and data visualizations were created using Microsoft PowerPoint (Microsoft 365, Redmond, WA, USA).

Ethical consideration
The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments and approved by the Ethics Committee of Xiangya Hospital (No. 2025030341). The requirement for informed consent was waived by the ethics board due to the retrospective design and the use of anonymized patient data.

Results
This real-world study initially identified 5,963 patients diagnosed with BC at the Department of Breast Surgery, Xiangya Hospital, Central South University, between 2015 and 2021. After excluding patients with non-hormone receptor-positive BC, DCIS, and specific invasive BC subtypes, a total of 2,180 patients with HR+ invasive BC were included for analysis. Based on the AJCC tumor staging system, 1,947 patients were classified as stage I–III and constituted the primary cohort for subsequent analyses. The patient selection flowchart is presented in Figure 1. Within this cohort, 437 patients (22.4%) were classified as stage I, 1,215 patients (62.4%) as stage II—including 798 patients (41.0%) with stage IIA and 417 patients (21.4%) with stage IIB—and 295 patients (15.2%) as stage III. Regarding nodal status, 1,128 patients (57.9%) were N0, 586 (30.1%) were N1, and 233 (11.5%) were classified as N2–3 (Table 1).
A total of 519 patients (26.7% of the cohort) met the eligibility criteria for the MonarchE trial. Among these, 367 patients fulfilled the criteria for Cohort 1, while 152 patients met the criteria for Cohort 2 (Table 2). In contrast, 1,130 patients (58.0%) were eligible for the NATALEE trial, more than twice the proportion of those meeting the MonarchE criteria. Of these, 295 patients were classified as stage III, 417 as stage IIB, 140 as stage IIA (N1), 78 as stage IIA (N0, G3), and 200 as stage IIA (N0, G2 with high-risk features) (Table 3). Figure 2 illustrates the distribution of patients meeting the inclusion criteria for both the MonarchE and NATALEE trials.
We compared the clinical characteristics of patients who met the eligibility criteria for the two trials. While patients eligible for the MonarchE trial generally exhibited higher risk profiles, the clinicopathological features of the NATALEE cohort were largely comparable to those of the MonarchE cohort. The median age was 49 years in the MonarchE group and 50 years in the NATALEE group. The proportion of patients aged ≤40 years was similar between the two groups (18.3% in MonarchE vs. 17.4% in NATALEE), whereas the NATALEE cohort included a higher percentage of patients aged >60 years (16.7%). The proportion of premenopausal patients exceeded 60% in both groups, with the remaining about 40% being predominantly postmenopausal. There was no statistically significant difference between the two groups in menopausal status. Regarding histologic grading, the MonarchE group had a greater proportion of grade 3 tumors (25.2%) and T3 stage disease (17.9%). In contrast, the NATALEE cohort showed a higher prevalence of patients with Ki-67 <20% (44.9%) (Table 4). In the NATALEE-eligible population, patients with N0 disease accounted for nearly one-quarter of the cohort (24.6%), whereas this subgroup was not included in the MonarchE trial. This distinction highlights a key difference in the eligibility criteria between the two studies.
In the subgroup of stage N0 patients enrolled in the NATALEE trial with high-risk factors, we applied the following inclusion criteria: T4N0, T3N0, or T2N0 with histologic grade G2 and Ki-67 ≥20%, histologic grade G3, or high genetic risk. We identified 312 patients who met these criteria, accounting for 27.6% of the NATALEE cohort. This subgroup had a similar age distribution to the overall cohort but exhibited a significantly higher proportion of patients with histologic grade G3 tumors (26.0%) and elevated Ki-67 levels (85.3%) (Table 5).

Discussion
This retrospective study analyzed a large Chinese cohort of HR+ EBC patients to evaluate eligibility and clinicopathological characteristics based on the MonarchE and NATALEE trial criteria. Our findings provide important insights into the real-world applicability of these trial-defined eligibility standards within the Chinese population.
The MonarchE and NATALEE trials have demonstrated the importance of adjuvant CDK4/6 inhibitors in high-risk HR+ EBC (12,13). However, our study highlights distinct differences in the populations eligible for these two trials within a real-world Asian context. Notably, 45.8% of NATALEE-eligible patients overlapped with the MonarchE criteria. The MonarchE-eligible cohort was predominantly node-positive, with higher-grade and stage disease compared with those eligible for NATALEE. Specifically, the proportion of grade 3 and T3 patients was greater in the MonarchE-eligible cohort (25.2% vs. 18.7% and 17.9% vs. 10.9%, respectively). Additionally, a higher percentage of MonarchE-eligible patients had a Ki-67 score of ≥20% (61.8% vs. 54.6%). These findings are consistent with the trial designs, as MonarchE focused on patients with more aggressive disease characteristics, whereas NATALEE included a broader population with intermediate risk. While most patients with early HR+/HER2− BC achieve favorable outcomes with standard therapy, a subset with high-risk features remain vulnerable to relapse and distant metastasis (12). For these patients, recurrence often leads to therapeutic challenges and poor prognosis. The MonarchE trial demonstrated that adding abemaciclib to ET significantly reduced recurrence risk in this high-risk group, improving both survival outcomes and quality of life. Our findings underscore the need to identify such high-risk subsets and tailor adjuvant strategies accordingly accurately.
In contrast, the NATALEE-eligible cohort was more heterogeneous, with a substantial proportion of N0 patients exhibiting high-risk biological features, including 27.6% of N0 patients in the NATALEE cohort. Recent data from specialized cancer centers or non-Asian registries corroborate our findings, reporting fewer patients eligible for adjuvant abemaciclib but a larger proportion eligible for adjuvant ribociclib (14-16). These results challenge the traditional nodal-centric risk stratification, emphasizing the need to consider tumor biology, such as Ki-67 and histologic grade, alongside nodal status when evaluating a patient’s risk.
A U.S. real-world study using the Concert AI Patient 360 electronic health record database, which included 7,060 patients enrolled between January 2015 and January 2023, reported that 2,163 (30.6%) and 1,023 (14.5%) patients met the NATALEE and MonarchE trial enrollment criteria, respectively (17). In contrast, our study found that the proportion of stage I patients was only 22.4%, significantly lower than the 60.4% reported in the U.S. real-world data. This discrepancy underscores the need for enhanced early screening for BC in the economically less-developed regions of central China. Based on clinical stage and biological characteristics, 58.0% of our cohort met the NATALEE trial criteria, while 26.7% met the MonarchE criteria. These figures were notably higher than those observed in the U.S. study, where 30.6% and 14.5% of patients met the respective trial criteria. A German real-world study showed that 43.0% of patients met the NATALEE trial inclusion criteria (14). These findings suggest that a significantly larger and more heterogeneous population in China may benefit from intensive CDK4/6 inhibitor therapy, with nearly 60% of patients potentially requiring adjuvant ribociclib to reduce recurrence risk.
In our cohort, the median age of patients meeting the NATALEE trial criteria was 50 years, with 17.4% of patients aged ≤40 years, significantly lower than the median age of 60 years reported in real-world studies from the United States and Europe (18,19). According to data from the YBCC-Catts study (20), the proportion of young BC patients under 35 years of age is significantly higher in China than in Western countries. Treatment with an aromatase inhibitor in combination with ovarian function suppression (OFS) has been shown to reduce both mortality and recurrence risk compared to tamoxifen therapy alone. Since the publication of the SOFT & TEXT studies (21), the use of OFS in premenopausal women has increased, with GnRH agonists (GnRHa) being the preferred choice for OFS in premenopausal HR+ EBC patients. Follow-up data from other studies have further confirmed that the use of OFS significantly reduces the risk of recurrence over 10 years (22,23), providing long-term survival benefits and enhancing the likelihood of a cure. In the NATALEE trial, OFS combined with aromatase inhibitors was used in premenopausal patients, whereas some patients in the MonarchE trial were treated with tamoxifen. Therefore, for HR+ premenopausal patients with medium- to high-risk BC, the addition of specific CDK4/6 inhibitors to GnRHa-based combination ET is recommended, while ensuring compatibility with the enrolled patient population in the relevant clinical studies.
During the study period, 1,130 patients met the NATALEE eligibility criteria, of whom 27.6% (312 cases) were N0. This proportion is higher than previously reported and was predominantly observed in patients with histologic grade 2 (G2) tumors and a Ki-67 index of ≥20% (24). A meta-analysis by the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) demonstrated that N0 BC patients remain at long-term risk of recurrence and metastasis, with recurrence risk increasing alongside tumor size and histologic grade (25). At ASCO 2023, updated data showed that adjuvant ribociclib significantly improved invasive disease-free survival (IDFS), distant disease-free survival, and distant recurrence-free survival (DRFS) in the N0 subgroup, with a safety profile comparable to that observed in the intention-to-treat population, providing valuable evidence for treating this patient subgroup (26). Moreover, during three years of adjuvant treatment with ribociclib 400 mg, patients exhibited good safety and tolerability, with no new safety signals identified and quality of life comparable to controls (13).
Although our study did not include direct survival or recurrence analyses due to incomplete follow-up data, evidence from external large-scale cohorts has consistently demonstrated the poor prognosis of patients fulfilling the NATALEE eligibility criteria. Data from the China National Cancer Center indicated that the 10-year recurrence rate in this population approached 40% (27). The recurrence or progression risk was approximately threefold higher in NATALEE-eligible patients compared with non-eligible counterparts [5-year real-world invasive disease-free survival (rwiDFS): 81.4% vs. 93.4%]. In the high-risk N0 subgroup, the risk was doubled (5-year rwiDFS: 85.9% vs. 93.4%), and even N1 patients without additional high-risk features still showed a 13.7% 5-year recurrence rate. Furthermore, a real-world analysis from the United Kingdom demonstrated that N0 patients with high-risk features had recurrence risks comparable to the overall N1 population (28). These findings underscore that despite advances in ET, recurrence remains a substantial and unmet clinical problem among patients meeting the NATALEE trial criteria. These findings suggest that in HR+/HER2− EBC patients clinically or pathologically diagnosed as N0, ribociclib treatment guided by NATALEE criteria may enhance efficacy without substantially increasing safety risks. Considering long-term tolerability and identifying a more precise benefit population, CDK4/6 inhibitors are expected to offer an optimized therapeutic strategy for HR+/HER2− EBC patients in China. In contrast, for the MonarchE eligibility criteria, several real-world studies from Western countries have confirmed similarly poor prognoses. A real-world study from the United States reported that, among patients meeting the MonarchE eligibility criteria, the risk of IDFS and DRFS events in cohort 1 was more than threefold higher compared with cohort 2 (29). A real-world study from Finland similarly demonstrated that patients fulfilling the cohort 1 criteria of the MonarchE trial had higher risks of recurrence, distant recurrence, and death compared with those in cohort 2. Specifically, the 5-year IDFS was 79.7% in cohort 1 vs. 89.3% in cohort 2, the 5-year DRFS was 82.4% vs. 92.9%, and the 5-year overall survival was 89.5% vs. 95.4%, respectively (30). These analyses consistently reported significantly higher recurrence risks and worse survival outcomes for patients meeting the MonarchE definition compared with lower-risk counterparts. However, existing data are limited to subgroup analyses within the MonarchE clinical trial, which demonstrated consistent benefit from abemaciclib plus ET in Chinese patients, but do not fully reflect outcomes in unselected real-world populations. In a clinical trial, the addition of abemaciclib to ET significantly improved both IDFS (85.9% vs. 79.1%) and DRFS (88.4% vs. 82.3%) compared with the control group (31). Considering potential differences in tumor biology, treatment patterns, and healthcare resources, this evidence gap highlights the urgent need for nationwide registry-based or multicenter real-world studies in China to validate and extend these findings.
Undoubtedly, several limitations exist in the present study. The retrospective, single-center design may restrict the generalizability of the findings. However, Xiangya Hospital’s position as a tertiary referral center helps mitigate potential selection bias. The absence of genetic testing data prevented a precise replication of the genetic risk criteria used in the NATALEE trial, necessitating the use of surrogate markers. Although Ki-67 is a widely used proliferative marker, it cannot fully substitute for multigene assays such as Oncotype DX or PAM50, which integrate broader molecular information. Relying on Ki-67 as a surrogate may therefore lead to misclassification bias, with some patients potentially overestimated or underestimated in terms of their true recurrence risk. This limitation could influence the comparability between MonarchE- and NATALEE-eligible populations and should be taken into account when interpreting our findings. Additionally, the lack of follow-up data on recurrence and survival outcomes precluded direct comparisons between the trial-eligible groups. Future prospective studies are essential to evaluate whether combining anatomic and biological risk factors (such as nodal status and Ki-67) can enhance prognostic accuracy beyond the trial-defined criteria. Moreover, cost-effectiveness analyses are urgently required to inform resource allocation, especially in regions with limited access to genetic testing or CDK4/6 inhibitors.

Conclusions
In this large real-world Chinese cohort, application of the NATALEE trial criteria identified a broader and more biologically diverse subset of patients with HR+/HER2− EBC, including a substantial proportion of N0 individuals with high-risk features. These results highlight the limitations of a nodal-centric approach and support a more nuanced risk-stratification strategy that integrates both anatomical staging and tumor biology. As adjuvant treatment paradigms continue to evolve, real-world evidence such as ours will be essential to inform clinical decision-making and promote the personalized use of CDK4/6 inhibitors in diverse patient populations.

Supplementary
The article’s supplementary files as