Development and validation of a survival-predicting nomogram for HER2-negative T1–3N0–1 breast cancer treated with breast-conserving surgery: a Surveillance, Epidemiology, and End Results (SEER) database analysis

Introduction

Breast carcinoma remains the predominant malignancy diagnosed among women worldwide and constitutes a leading cause of cancer-associated mortality (1). Breast-conserving surgery (BCS) combined with radiotherapy is the cornerstone of treatment for early-stage breast cancer (BC), offering excellent oncologic control and preserved quality of life for patients with T1–3N0–1 disease (2). Within this population, human epidermal growth factor receptor 2 (HER2)-negative tumors represent the most common molecular subtype, yet exhibit significant biological heterogeneity, leading to variable clinical outcomes (3,4). While patients with low-risk features may derive minimal benefit from systemic chemotherapy, those with high-risk clinicopathological factors—such as larger tumor size (T2–3), high histological grade, elevated Ki-67 index, or limited lymph node involvement—face a substantially increased risk of recurrence and mortality, making chemotherapy a critical component of their multimodal management (5-7).

Importantly, the efficacy of chemotherapy, particularly when administered preoperatively (neoadjuvant chemotherapy; NAC), varies markedly across this spectrum. Patients achieving a complete response (CR) following NAC demonstrate significantly improved long-term survival compared to those with residual disease (non-CR) (8,9). This prognostic dichotomy is especially pronounced in aggressive HER2-negative subtypes like triple-negative breast cancer (TNBC), where CR is strongly associated with excellent event-free survival (10,11). Conversely, failure to achieve CR, particularly in high-risk subgroups, identifies a cohort with persistent micrometastatic burden and a heightened need for intensified adjuvant strategies or novel therapeutic approaches (12). Furthermore, the degree of residual disease burden post-NAC provides refined prognostic stratification beyond the binary CR/non-CR distinction (13).

Despite standardized surgical and radiation approaches, the inability to consistently predict which patients will derive maximal benefit from chemotherapy or achieve CR hampers optimal personalized treatment selection for HER2-negative T1–3N0–1 BC patients undergoing BCS. Consequently, robust tools integrating baseline risk factors and dynamic response assessments are urgently needed to guide therapeutic decision-making, optimize survival outcomes, and spare low-risk patients from unnecessary toxicity. Several prognostic models and nomograms have indeed been developed for HER2-negative BC, including those utilizing the Surveillance, Epidemiology, and End Results (SEER) database. For instance, existing models predict survival in hormone receptor (HR)-positive/HER2-negative patients with axillary lymph node metastasis or guide adjuvant chemotherapy decision. These contributions are valuable; however, a systematic consideration of the literature reveals specific limitations in the context of our target population (14). Many existing models either focus on all surgical types collectively or are not specifically designed and validated for the distinct T1–3N0–1 population undergoing BCS. Furthermore, few integrate the dynamic assessment of treatment response with baseline tumor biology and sociodemographic factors into a single, practical tool for this specific cohort.

Consequently, a critical gap remains in consistently predicting which patients with HER2-negative T1–3N0–1 BC undergoing BCS will derive maximal benefit from chemotherapy or achieve a pathological complete response. This hampers optimal personalized treatment selection. Therefore, robust tools that uniquely combine baseline risk stratification with dynamic response assessments are needed to guide therapeutic decision-making, optimize survival outcomes, and spare low-risk patients from unnecessary toxicity.

To address this gap, our study aims to develop and validate prognostic nomograms for overall survival (OS) and cancer-specific survival (CSS) specifically within this well-defined cohort. The novelty of our model lies in its unique combination of predictors—integrating baseline clinicopathological variables, treatment response to NAC, and sociodemographic factors—tailored expressly for HER2-negative T1–3N0–1 patients treated with BCS, thereby facilitating more personalized adjuvant therapy decisions. We present this article in accordance with the TRIPOD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-301/rc).

Methods

Data collection and patient selection

Based on the SEER Research Plus Data 22 registry [2000–2019], we initially identified female patients diagnosed with BC. To align with the consistent application of the 7th edition American Joint Committee on Cancer (AJCC) staging criteria, the study period was restricted to cases diagnosed between January 2010 and December 2016. Key inclusion criteria were: (I) HER2-negative status; (II) treatment with BCS followed by postoperative radiotherapy; and (III) tumor, node, metastasis (TNM) stage T1–3N0–1M0. Eligible cases were randomly split into training and internal validation sets at a 7:3 ratio. For each patient, demographic, clinicopathological, and treatment-related variables were extracted, including age at diagnosis, race, tumor laterality, AJCC TNM stage, histological grade, primary site, histology, lymph node involvement, estrogen receptor (ER) status, progesterone receptor (PR) status, HER2 status, tumor size, BC subtype, and sequence of tumor occurrence. Information on neoadjuvant therapy receipt, cause-specific death classification, and survival months (with a minimum of >0 days) was also collected. Patients with incomplete survival or follow-up records, or with zero or missing survival time, were excluded from the final analysis. The resulting cohort enabled robust evaluation of long-term survival and identification of independent prognostic factors. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Variable selection

The analysis incorporated the following variables: age at diagnosis, race, laterality, TNM stage, grade, tumor location, histological type, lymph node metastasis, BC subtype, tumor size, ER status, PR status, HER2 status, cause-specific death, tumor sequence, and response to neoadjuvant therapy. In accordance with the SEER database classification, several variables were consolidated into specific categories for analysis.

Age was categorized into two groups: <45 years and ≥70 years. Race was grouped as White and other. The primary tumor site was divided into five categories: lower-inner, lower-outer, upper-inner, upper-outer, and others. Tumor grade was classified as I/II versus III/IV. Histological type included infiltrating duct carcinoma and others. Laterality was designated as left or right. Marital status was grouped as married or unmarried. Chemotherapy use was classified as no/unknown or yes. Both PR and ER status were defined as negative or positive. The sequence number of tumors was categorized as one primary only or more than one primary. Tumor stage was classified into stages I, II, and IIIA. T stage was grouped as T1, T2, and T3, while N stage was categorized as N0 and N1.

Furthermore, a CR was defined as the absence of invasive cancer in both the breast (ypT0/Tis) and lymph nodes (ypN0) in the pathologic assessment following NAC, as recorded in the corresponding SEER fields. Any other outcome was classified as “non-CR”. A “single primary tumor” (one primary only) was defined as the first and only malignant tumor in the patient’s lifetime, as per the SEER sequence number code ‘00’. Any code indicating multiple tumors was categorized accordingly.

Statistical analysis

Prognostic factors were identified through univariate Cox proportional hazards regression and then assessed in a multivariate model. This analysis provided hazard ratios along with 95% confidence intervals (CIs) for each variable, ensuring robust identification of independent predictors.

Based on the multivariate Cox regression model implemented in R, we constructed and validated nomograms to estimate 1-, 3-, and 5-year OS and CSS. Model discrimination was assessed using the concordance index (C-index) and time‑dependent receiver operating characteristic (ROC) curves, with performance further quantified by the area under the curve (AUC). Calibration plots compared predicted versus observed survival. Clinical utility was assessed via decision curve analysis (DCA), quantifying net benefits across threshold probabilities. The development cohort was segmented into risk groups based on total points, and survival differences were analyzed using the Kaplan-Meier method with log-rank tests. Propensity score matching (PSM) was applied to evaluate clinical intervention effects on outcomes. Statistical analyses used R (v3.6.1) and SPSS (v25.0; IBM). The “rms”, “survival”, “magick”, “timeROC”, “ggplotify”, and “cowplot” R packages facilitated nomogram development and validation. Statistical significance was defined as P<0.05. All P values are from two-sided tests.

Results

Baseline clinical features

In summary, this study included 8,384 operable BC patients, categorized into a training cohort of 5,869 and an internal test cohort of 2,515 (Figure S1). The majority of patients (77.8% training, 78.1% test) were aged ≥45 years. All patients underwent BCS, with approximately two-thirds receiving chemotherapy (64.2% training, 65.4% test). Most tumors were moderately/well differentiated (grade I/II: 66.0% training, 66.3% test), while poorly differentiated/undifferentiated tumors (grade III/IV) accounted for 34.0% and 33.7% respectively. Infiltrating duct carcinoma was the predominant histologic subtype (77.0% training, 77.4% test). The HR⁺/HER2⁻ subtype represented the vast majority of cases (85.2% training, 86.7% test), with ER positivity observed in 84.6% and 85.7% of cases, and PR positivity in 74.7% and 75.9% across cohorts. For tumor staging, stage II was most frequent (62.7% training, 63.5% test), with T2 being the commonest T stage (41.5% both cohorts) and N1 predominating in nodal status (51.3% training, 53.5% test). Baseline characteristics of the study population are summarized in Table 1. Pearson’s Chi-squared tests indicated no statistically significant differences in the distribution of clinicopathological features between the training and validation cohorts (all P>0.05).

Variable feature importance of survival prediction

We systematically evaluated prognostic factors for OS and CSS in the training cohort. Candidate variables from univariate Cox regression analyses were integrated into multivariate models to identify independent predictors, with full results detailed across Tables 2-5.

For OS, multivariate analysis revealed that older age (≥45 years: hazard ratio =1.33; 95% CI: 1.09–1.61), higher tumor grade (III/IV: hazard ratio =1.71; 95% CI: 1.43–2.05), unmarried status (hazard ratio =1.62; 95% CI: 1.40–1.88), advanced T stage (T2: hazard ratio =1.72, 95% CI: 1.29–2.29; T3: hazard ratio =1.90, 95% CI: 1.34–2.68) were independent risk factors. Conversely, ER positivity (hazard ratio =0.72; 95% CI: 0.57–0.92), PR positivity (hazard ratio =0.67; 95% CI: 0.54–0.83), single primary tumor (hazard ratio =0.56; 95% CI: 0.47–0.66) and receipt of chemotherapy (hazard ratio =0.59; 95% CI: 0.49–0.70) were significantly associated with improved OS.

For CSS, independent risk factors included higher tumor grade (III/IV: hazard ratio =1.70; 95% CI: 1.35–2.14), unmarried status (hazard ratio =1.38; 95% CI: 1.14–1.67), advanced T stage (T2: hazard ratio =2.23; 95% CI: 1.48–3.37; T3: hazard ratio =3.28; 95% CI: 1.92–5.60) and N1 stage (hazard ratio =1.31; 95% CI: 1.00–1.70). Protective factors were ER positivity (hazard ratio =0.70; 95% CI: 0.52–0.94), PR positivity (hazard ratio =0.55; 95% CI: 0.42–0.71) and single primary tumor (hazard ratio =0.57; 95% CI: 0.46–0.71).

Nomogram construction

Based on the multivariate Cox analysis of the training cohort, we constructed nomograms to predict OS and CSS (Figures 1,2). Each independent prognostic factor was assigned a weighted score (range, 0–100) proportional to its contribution to the outcome. The total point score for an individual, obtained by summation of all factor scores, was then converted into estimated probabilities for 1-, 3-, and 5-year OS and CSS. A higher total score consistently indicated a poorer prognosis.

Figure 1 Nomogram for OS prediction in the patients. ER, estrogen receptor; OS, overall survival; PR, progesterone receptor; T, tumor.

Figure 2 Nomogram for CSS prediction in the patients. CSS, cancer-specific survival; ER, estrogen receptor; N, node; PR, progesterone receptor; T, tumor.

Model validation

In the training cohort, the OS nomogram demonstrated a significantly higher C-index than the AJCC 7th staging system [0.69 (95% CI: 0.67–0.71) vs. 0.63 (95% CI: 0.61–0.65)]. Time‑dependent ROC analysis yielded AUCs of 0.684 (1-year), 0.701 (3-year), and 0.697 (5-year) (Figure 3A). Similarly, for CSS, the nomogram achieved a C-index of 0.74 (95% CI: 0.72–0.76) in the training cohort, with corresponding 1-, 3-, and 5-year AUCs of 0.797, 0.780, and 0.748 (Figure 3B).

Figure 3 ROC curves of the nomogram for 1-, 3-, and 5-year OS (A) and CSS (B) in training set. AUC, area under the curve; CSS, cancer-specific survival; OS, overall survival; ROC, receiver operating characteristic.

Validation results confirmed the robustness of both models. For OS, the validation cohort showed a Cindex of 0.67 (95% CI: 0.64–0.71) and 1-/3-/5-year AUCs of 0.672, 0.718, and 0.687 (Figure S2A). For CSS, the validation C-index was 0.77 (95% CI: 0.73–0.80) with AUCs of 0.813, 0.832, and 0.778 (Figure S2B).

Calibration plots indicated excellent agreement between predicted and observed outcomes for OS in both the training (Figure 4A-4C) and validation sets (Figure 4D-4F). High concordance was also observed for CSS in the training (Figure S3A-S3C) and validation cohorts (Figure S3D-S3F).

Figure 4 Calibration plots for 1-, 3-, and 5-year OS. (A) Training set, 1-year OS. (B) Training set, 3-year OS. (C) Training set, 5-year OS. (D) Validation set, 1-year OS. (E) Validation set, 3-year OS. (F) Validation set, 5-year OS. OS, overall survival.

DCA revealed favourable clinical utility of the OS nomogram in the training (Figure 5A-5C) and validation cohorts (Figure 5D-5F). Similarly, the CSS nomogram showed valuable clinical net benefit in both the training (Figure S4A-S4C) and validation sets (Figure S4D-S4F).

Figure 5 Decision curve analysis of nomogram for the 1-, 3-, and 5-year OS prediction in training cohort (A-C) and validation cohort (D-F). None: none of the patients have a bad outcome. All: bad outcomes occur in all patients. OS, overall survival.

Therapeutic efficacy across risk-stratified subgroups

Drawing upon X-tile-derived optimal cut-off values, patients were stratified into three distinct prognostic tiers: low-risk (total score: OS <129, CSS <104), intermediate-risk (OS: 129–247, CSS: 104–205), and high-risk (OS ≥248, CSS ≥206) (Figure S5). Kaplan-Meier survival curves demonstrated statistically significant stratification across these groups for both OS and CSS (Figure 6), with superior outcomes observed in the low-risk cohort and progressively poorer survival associated with higher risk scores.

Figure 6 Kaplan-Meier OS (A) and CSS (B) curves with different risk group stratified by the nomogram. CSS, cancer-specific survival; OS, overall survival.

Subgroup analyses, performed after PSM, evaluated the differential impact of systemic therapies. In both the low- and intermediate-risk strata, neither adjuvant chemotherapy nor NAC conferred a significant survival advantage for OS or CSS (Figure 7 and Figure S6). However, among patients receiving NAC in the low- and medium-risk groups, those achieving CR demonstrated a significantly better prognosis than those with NCR (Figure 8). What is more, we found that both adjuvant chemotherapy and NAC demonstrated improved OS and CSS in high-risk group patients (Figure 9). Interestingly, patients achieving a CR after NAC had superior survival outcomes compared to those receiving postoperative chemotherapy, while patients with NCR to NAC showed inferior survival outcomes compared to postoperative chemotherapy patients (Figure 9). It was revealed that within the high-risk group, patients achieving a CR to NAC had a significantly better prognosis than those with an NCR (Figure S7).

Figure 7 Comparison of Kaplan-Meier survival curves between patients who did and did not receive adjuvant chemotherapy in different risk groups. (A) OS in the low-risk group. (B) CSS in the low-risk group. (C) OS in the middle-risk group. (D) CSS in the middle-risk group. CSS, cancer-specific survival; OS, overall survival.

Figure 8 Comparison of survival outcomes between patients achieving a CR and those with an NCR following neoadjuvant chemotherapy, stratified by risk group. (A) OS in the low-risk group. (B) CSS in the low-risk group. (C) OS in the intermediate-risk group. (D) CSS in the intermediate-risk group. CR, complete response; CSS, cancer-specific survival; NCR, non-complete response; OS, overall survival.

Figure 9 Kaplan-Meier analysis for different chemotherapy strategies in high-risk patients group. Adjuvant vs. no chemotherapy in high-risk patients: OS (A) and CSS (E). Neoadjuvant vs. no chemotherapy in high-risk patients: OS (B) and CSS (F). The comparison between patients who received adjuvant chemotherapy and those who receive neoadjuvant chemotherapy in the high-risk patients group of CR for OS (C) and CSS (G). The comparison between patients who received adjuvant chemotherapy and those who receive neoadjuvant chemotherapy in the high-risk patients group of NCR for OS (D) and CSS (H). CR, complete response; CSS, cancer-specific survival; NCR, non-complete response; OS, overall survival.

Discussion

BC remains a leading cause of cancer-related mortality in women worldwide, with HER2-negative subtypes constituting over 70% of early-stage cases (15). BCS combined with radiotherapy is the cornerstone of treatment for T1–3N0–1 disease, offering excellent locoregional control while preserving quality of life (16). However, significant heterogeneity exists within this population, where patients with adverse features—including larger tumor size (T2–3), high histological grade, limited nodal involvement (N1), or hormone receptor negativity—experience up to a 3-fold increase in recurrence risk compared to their low-risk counterparts (17). Critically, the benefit of systemic chemotherapy, particularly when administered NAC, varies substantially across this spectrum. Pathological CR to NAC is a well-established surrogate for long-term survival in aggressive subtypes like TNBC, yet fewer than 40% of HER2-negative patients achieve this milestone (18). Conversely, residual disease signifies persistent micrometastatic burden and necessitates intensified adjuvant strategies (19). Despite these insights, the inability to reliably predict chemotherapy response or identify patients most likely to benefit from NAC hampers personalized decision-making (20,21). Our study addresses this unmet need by developing and validating the first nomograms integrating baseline clinicopathological factors and treatment response to predict long-term survival in HER2-negative T1–3N0–1 BC patients treated with BCS.

Through rigorous analysis of 8,384 patients from the SEER registry, we identified tumor grade, T stage, ER/PR status, marital status, and tumor sequence as independent predictors of OS and CSS. Advanced T stage (T2: OS hazard ratio =1.72, CSS hazard ratio =2.23; T3: OS hazard ratio =1.90, CSS hazard ratio =3.28) and high-grade histology (grade III/IV: OS hazard ratio =1.71, CSS hazard ratio =1.70) emerged as dominant risk factors, aligning with established literature on their association with aggressive tumor biology and metastatic potential (22). The protective role of hormone receptor positivity (ER⁺: OS hazard ratio =0.72; PR⁺: OS hazard ratio =0.67) underscores the therapeutic advantage of endocrine therapies in this population—a finding consistent across previous studies (23,24). Notably, unmarried status independently predicted poorer survival (OS hazard ratio =1.62, CSS hazard ratio =1.38), likely reflecting socioeconomic disparities in treatment access, adherence, or social support systems rather than biological mechanisms. Most innovatively, our study established “single primary tumor” as a novel favorable prognostic factor (OS hazard ratio =0.56; CSS hazard ratio =0.57). This finding challenges conventional staging paradigms and suggests that multiple primaries may indicate field cancerization, genetic predisposition, or compromised immune surveillance—mechanisms warranting further molecular investigation. These variables were integrated into visually accessible nomograms that outperform the AJCC 7th staging system, providing clinicians with a dynamic tool for risk quantification beyond static anatomical staging.

A pivotal contribution of this work lies in its risk-adapted evaluation of chemotherapy efficacy. Using X-tile-derived thresholds, we stratified patients into low-, middle-, and high-risk groups. Crucially, adjuvant chemotherapy provided no significant OS or CSS benefit in low-risk patients, while in middle-risk groups, its impact was marginal. This aligns with recent de-escalation trials demonstrating that chemotherapy may be safely omitted in genomically low-risk hormone receptor-positive disease (25,26). Our findings extend this paradigm to clinicopathologically defined low-risk HER2-negative BC, suggesting that routine chemotherapy exposes these patients to unnecessary toxicity without survival gains.

Conversely, high-risk patients derived substantial benefit from chemotherapy. More importantly, we observed a differential effect based on treatment timing and response. NAC significantly improved survival only when CR was achieved. In middle-risk patients, CR after NAC yielded superior outcomes versus NCR. High-risk patients achieving CR with NAC had better OS and CSS than those receiving adjuvant chemotherapy, whereas NCR patients fared worse than their adjuvant-treated counterpart.

These results illuminate two critical principles. CR is a transformative endpoint—its achievement identifies chemosensitive tumors where NAC may be curative, particularly in high-risk disease. NCR requires alternative strategies—residual disease signals intrinsic resistance, necessitating postoperative escalation or novel agents. This paradigm shift toward response-adapted therapy underscores NAC’s dual role: tumor downstaging and “in vivo” chemosensitivity testing. Our data suggest that high-risk patients should be prioritized for NAC, with subsequent treatment intensity tailored to pathological response.

While the nomograms demonstrated a statistically significant improvement over the AJCC 7th system, their discriminative ability (C-index: 0.69 for OS, 0.74 for CSS) must be interpreted as modest. Consequently, these models are best suited for group-level risk stratification and should serve as an adjunct to, rather than a replacement for, clinical judgment in individual patient management. Calibration showed excellent concordance in predicted group and observed group. DCA confirmed net benefit across threshold probabilities. These tools address a critical gap in managing HER2-negative BC. By quantifying individualized risk, clinicians can avoid overtreatment in low-risk patients (33% of cohort), sparing toxicity. High-risk patients (27% of the cohort) should be prioritized for NAC to capitalize on the survival gains linked to achieving CR. Patients who do not achieve CR can then be redirected to adjuvant trials exploring novel agents.

Several limitations merit acknowledgment: (I) the retrospective SEER data lack granular details on chemotherapy regimens, radiation doses, endocrine therapy adherence, and molecular markers; (II) CR assessment was registry-based without central pathology review, potentially misclassifying residual disease burden; and (III) using “unmarried” status as a risk factor may mask the interplay of biological, social, and healthcare-access variables. External validation in prospective cohorts is essential. It is important to acknowledge that while our nomogram demonstrated a statistically significant improvement over the AJCC 7th system, the absolute gain in discriminative ability (C-index) is modest. This inherently limits its precision for predicting outcomes for individual patients. Therefore, the primary clinical value of this model lies not in providing definitive predictions, but in serving as a decision aid for risk stratification at the group level, as supported by the net benefit shown in DCA. Future efforts should therefore focus on integrating genomic signatures with clinicopathological nomograms to refine risk prediction, exploring biomarkers of NAC response for early identification of non-responders, and designing trials that randomize high-risk patients to NAC versus adjuvant chemotherapy with response-adaptive crossover.

Conclusions

In HER2-negative early-stage BC treated with BCS, our study delivers three key advances: first, we have developed validated nomograms integrating tumor biology, treatment response, and social factors that predict long-term survival more accurately than conventional staging. Second, a risk-adapted strategy, in which chemotherapy avoidance may be considered for low-risk groups, NAC could be prioritized for high-risk patients, and alternative strategies might be explored for NAC non-responders. Third, we identified “single primary tumor” status as a novel prognostic indicator, suggesting unexplored biological mechanisms in tumorigenesis. These tools empower clinicians to navigate HER2-negative BC heterogeneity—optimizing survival while minimizing unnecessary toxicity. Prospective validation and biomarker integration will further advance precision medicine for this population.

However, these findings and their implications must be interpreted in the context of the study’s limitations, including its retrospective nature and the need for external validation in independent, prospective cohorts. If validated, these tools could eventually help clinicians navigate the heterogeneity of HER2-negative BC—with the goal of optimizing survival while minimizing unnecessary toxicity. Therefore, prospective validation and the integration of novel biomarkers are essential next steps to confirm the clinical utility of our model and advance precision medicine for this population.

Acknowledgments

We would like to thank all the researchers for the SEER program.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-301/rc

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-301/prf

Funding: This study was supported by the National Natural Science Foundation of China (Nos. 82372078 and 82203688).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-301/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2024;74:229-63. [Crossref] [PubMed]
2. Adesunkanmi AO, Wuraola FO, Fagbayimu OM, et al. Oncoplastic Breast-Conserving Surgery in African Women: A Systematic Review. JCO Glob Oncol 2024;10:e2300460. [Crossref] [PubMed]
3. Hortobagyi GN, Lacko A, Sohn J, et al. A phase III trial of adjuvant ribociclib plus endocrine therapy versus endocrine therapy alone in patients with HR-positive/HER2-negative early breast cancer: final invasive disease-free survival results from the NATALEE trial. Ann Oncol 2025;36:149-57. [Crossref] [PubMed]
4. Hamilton E, Oliveira M, Turner N, et al. A phase I dose escalation and expansion trial of the next-generation oral SERD camizestrant in women with ER-positive, HER2-negative advanced breast cancer: SERENA-1 monotherapy results. Ann Oncol 2024;35:707-17. [Crossref] [PubMed]
5. Cardoso F, O'Shaughnessy J, Liu Z, et al. Pembrolizumab and chemotherapy in high-risk, early-stage, ER(+)/HER2(-) breast cancer: a randomized phase 3 trial. Nat Med 2025;31:442-8. [Crossref] [PubMed]
6. Michel LL, Schwarz D, Romar P, et al. Efficacy of Hand Cooling and Compression in Preventing Taxane-Induced Neuropathy: The POLAR Randomized Clinical Trial. JAMA Oncol 2025;11:408-15. [Crossref] [PubMed]
7. Foidart P, Jerusalem G. Is adjuvant olaparib standard of care in all high-risk HER2-negative early breast cancer presenting gBRCA1/2 mutations? Gland Surg 2025;14:257-62. [Crossref] [PubMed]
8. Villacampa G, Navarro V, Matikas A, et al. Neoadjuvant Immune Checkpoint Inhibitors Plus Chemotherapy in Early Breast Cancer: A Systematic Review and Meta-Analysis. JAMA Oncol 2024;10:1331-41. [Crossref] [PubMed]
9. Loi S, Salgado R, Curigliano G, et al. Neoadjuvant nivolumab and chemotherapy in early estrogen receptor-positive breast cancer: a randomized phase 3 trial. Nat Med 2025;31:433-41. [Crossref] [PubMed]
10. Connors C, Valente SA, ElSherif A, et al. Real-World Outcomes with the KEYNOTE-522 Regimen in Early-Stage Triple-Negative Breast Cancer. Ann Surg Oncol 2025;32:912-21. [Crossref] [PubMed]
11. Ligorio F, Vingiani A, Torelli T, et al. Early downmodulation of tumor glycolysis predicts response to fasting-mimicking diet in triple-negative breast cancer patients. Cell Metab 2025;37:330-344.e7. [Crossref] [PubMed]
12. Dugo M, Huang CS, Egle D, et al. The Immune-Related 27-Gene Signature DetermaIO Predicts Response to Neoadjuvant Atezolizumab plus Chemotherapy in Triple-Negative Breast Cancer. Clin Cancer Res 2024;30:4900-9. [Crossref] [PubMed]
13. Shien T, Tsuda H, Sasaki K, et al. Comparison of proportions and prognostic impact of pathological complete response between evaluations of representative specimen and total specimen in primary breast cancer after neoadjuvant chemoradiotherapy: an ancillary study of JCOG0306. Breast Cancer Res Treat 2024;208:145-54. [Crossref] [PubMed]
14. Hu Z, Wang W, Chen Y, et al. Development and validation of a radiomics-based nomogram for predicting two subtypes of HER2-negative breast cancer. Gland Surg 2024;13:2300-12. [Crossref] [PubMed]
15. Giaquinto AN, Sung H, Newman LA, et al. Breast cancer statistics 2024. CA Cancer J Clin 2024;74:477-95. [Crossref] [PubMed]
16. Chen W, Zhang Y, Zhang L, et al. Intraoperative evaluation of tumor margins using a TROP2 near-infrared imaging probe to enable human breast-conserving surgery. Sci Transl Med 2024;16:eado2461. [Crossref] [PubMed]
17. Geyer CE Jr, Untch M, Huang CS, et al. Survival with Trastuzumab Emtansine in Residual HER2-Positive Breast Cancer. N Engl J Med 2025;392:249-57. [Crossref] [PubMed]
18. Huang YH, Shi ZY, Zhu T, et al. Longitudinal MRI-Driven Multi-Modality Approach for Predicting Pathological Complete Response and B Cell Infiltration in Breast Cancer. Adv Sci (Weinh) 2025;12:e2413702. [Crossref] [PubMed]
19. Wang X, Chen B, Zhang H, et al. Integrative analysis identifies molecular features of fibroblast and the significance of fibrosis on neoadjuvant chemotherapy response in breast cancer. Int J Surg 2024;110:4083-95. [Crossref] [PubMed]
20. Chen Y, Huang Y, Li W, et al. Intratumoral microbiota-aided fusion radiomics model for predicting tumor response to neoadjuvant chemoimmunotherapy in triple-negative breast cancer. J Transl Med 2025;23:352. [Crossref] [PubMed]
21. Qiu C, Wei Y, Li J. Can negative axillary ultrasound reliably predict pathologically negative axillary lymph node status in breast cancer patients with cT ≤3 cm, cN0, and HER2-positive?-a retrospective, single-institution study. Gland Surg 2024;13:1511-21. [Crossref] [PubMed]
22. Wu Y, Zhang Y, Duan S, et al. Survival prediction in second primary breast cancer patients with machine learning: An analysis of SEER database. Comput Methods Programs Biomed 2024;254:108310. [Crossref] [PubMed]
23. Huo M, Zhang J, Hou M, et al. Development and Validation of a Nomogram for Prognosis Prediction in Patients with Synchronous Primary Thyroid and Breast Cancer Based on SEER Database. Cancer Invest 2024;42:212-25. [Crossref] [PubMed]
24. Qiu Y, Chen Y, Shen H, et al. Triple-negative breast cancer survival prediction: population-based research using the SEER database and an external validation cohort. Front Oncol 2024;14:1388869. [Crossref] [PubMed]
25. Corti C, Binboğa Kurt B, Koca B, et al. Estrogen Signaling in Early-Stage Breast Cancer: Impact on Neoadjuvant Chemotherapy and Immunotherapy. Cancer Treat Rev 2025;132:102852. [Crossref] [PubMed]
26. Bhimani J, O'Connell K, Ergas IJ, et al. Trends in chemotherapy use for early-stage breast cancer from 2006 to 2019. Breast Cancer Res 2024;26:101. [Crossref] [PubMed]