Development and validation of nomograms predicting survival in female patients with HER2-positive T1–3N0–1 breast cancer following breast-conserving surgery: a Surveillance, Epidemiology, and End Results database analysis

Introduction

Breast cancer (BC) remains the most prevalent malignancy and a leading cause of cancer-related mortality among women worldwide (1). For early-stage patients (T1–3N0–1), breast-conserving surgery (BCS) followed by radiotherapy represents the standard of care, balancing oncological efficacy with quality-of-life preservation (2). Within this cohort, human epidermal growth factor receptor 2 positive (HER2⁺) disease—accounting for 15–20% of BC cases—exhibits distinct biological aggression but has witnessed transformative survival gains through anti-HER2 targeted therapies (3,4). Nevertheless, significant outcome heterogeneity persists, with 5-year recurrence rates ranging from 10% to >30% despite modern treatments (5).

The integration of neoadjuvant chemotherapy (NAC) with dual HER2 blockade has become a cornerstone for stage II–III HER2⁺ BC, enabling pathological complete response (CR) rates of 40–70% (6). Achieving CR strongly correlates with superior event-free and overall survival (OS), while residual disease identifies high-risk patients warranting treatment escalation (7,8). However, critical knowledge gaps persist in T1–3N0–1 patients undergoing BCS: (I) baseline predictors of survival independent of NAC response remain poorly defined; (II) the differential benefit of chemotherapy sequencing (NAC vs. adjuvant) across molecularly distinct subgroups is unquantified; and (III) conventional tumor, node and metastasis (TNM) staging inadequately stratifies risk in this population (9,10).

This prognostic uncertainty hampers personalized management. While high-risk patients may benefit from NAC’s dual advantages of tumor downstaging and biological prognostication, low-risk subgroups might experience unnecessary toxicity from intensive regimens (11,12). Conversely, under-treatment in chemo-resistant phenotypes persists due to the lack of validated tools identifying candidates for novel adjuvant strategies. Existing HER2⁺ prognostic models frequently omit key clinicodemographic variables and fail to integrate dynamic response assessment (13).

To address these gaps, we leveraged the Surveillance, Epidemiology, and End Results (SEER) database to develop and validate the first nomograms predicting OS and cancer-specific survival (CSS) specifically for HER2⁺ T1–3N0–1 BC patients treated with BCS. Our aims were threefold: (I) identify independent baseline prognosticators; (II) construct visual tools integrating clinicopathological and treatment-response variables; and (III) stratify patients by risk to quantify differential chemotherapy benefits. This approach addresses an unmet need for precision in early HER2⁺ BC management, in which optimized therapy intensity must balance survival against toxicity. We present this article in accordance with the TRIPOD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-2025-302/rc).

Methods

Data collection and patient selection

Patients with BC were identified from the SEER Research Plus Data 22 registry [2010–2016]. To ensure staging consistency using the 7th American Joint Committee on Cancer (AJCC) system, we restricted analysis to cases diagnosed between January 2010 and December 2016. Inclusion criteria comprised: (I) HER2 positive; (II) completion of BCS with postoperative radiotherapy; (III) TNM stage T1–3N0–1M0. This cohort enabled analysis of survival rates and prognostic factors.

Each patient’s comprehensive information encompassed a range of age, race, laterality, TNM stage, grade, location, histological type, lymph node metastasis, breast cancer (BC) subtype, tumor size, estrogen receptor (ER) status, progesterone receptor (PR) status, HER2 status, cause-specific death, tumor sequence, response to neoadjuvant therapy and survival months (more than 0 days of survival).

Patients were randomly allocated to training and validation cohorts (7:3 ratio) to ensure analytical robustness. Exclusion criteria comprised: (I) incomplete survival/follow-up data; (II) zero survival time or missing survival data. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Statistical analysis

Through the utilization of Cox regression models, we performed calculations to determine a 95% confidence interval (CI) and hazard ratio (HR). In order to identify potential prognostic factors, those showing significant differences in the univariate Cox regression analysis were further examined through multivariate analysis. What’s more, univariate and multivariate analyses based on the competing risks model were performed.

Using R software, we developed multivariate analysis-based nomograms to predict 3-, 5-, and 7-year OS and CSS. Model discrimination was evaluated using the concordance index (C-index), time-dependent receiver operating characteristic (ROC) curves, and 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). Multiple imputation (MI) by Chained Equations was performed to handle missing data in the SEER database and enhance the robustness of the statistical analysis. The “rms”, “survival”, “magick”, “timeROC”, “ggplotify”, and “cowplot” R packages facilitated nomogram development and validation. Statistical significance was defined as P<0.05.

Results

Baseline clinical features

This study included a total of 13,875 operable BC patients, stratified into a training cohort (n=9,713) and an internal test cohort (n=4,162) (Figure S1). The majority of patients were aged ≥50 years (80.0% in training, 79.2% in test). Infiltrating duct carcinoma was the predominant histological subtype in both cohorts (93.5% training, 93.3% test), and most tumors were classified as poorly differentiated/undifferentiated (grade III/IV: 53.2% training, 54.4% test). The hormone receptor-positive (HR⁺)/HER2⁺ BC subtype represented the largest proportion (75.4% training, 75.5% test). Correspondingly, ER positivity was observed in 74.0% (training) and 73.8% (test) of cases, while PR positivity occurred in 57.1% and 58.8% of cases, respectively. Chemotherapy was administered to 76.9% (training) and 77.1% (test) of patients. Regarding tumor staging, stage I (TNM) was most frequent (58.3% training, 56.9% test), T1 tumors predominated in T-stage (66.3% training, 63.6% test), and N0 represented the majority of nodal statuses (77.9% training, 78.4% test). Comparative analysis using Pearson’s χ² tests revealed no statistically significant differences in the distributions of most baseline characteristics between the training and test cohorts (all P>0.05), with the exceptions of T-stage (P=0.002) and sequence number (indicating multiple primaries, P=0.03). The full baseline characteristics are detailed in Table 1.

Variable feature importance of survival prediction

For OS, univariate Cox regression in the training cohort (n=9,713) identified significant associations between survival outcomes and multiple variables: advanced age (≥50 years: HR =2.40, 95% CI: 1.85–3.12), White race (HR =1.30, 95% CI: 1.06–1.60), non-infiltrating duct carcinoma (HR =0.63, 95% CI: 0.42–0.93), chemotherapy (HR =0.49, 95% CI: 0.42–0.57), ER positive (HR =0.73, 95% CI: 0.62–0.87), PR positive (HR =0.76, 95% CI: 0.65–0.89), one primary only (HR =0.37, 95% CI: 0.31–0.43), unmarried status (HR =1.90, 95% CI: 1.62–2.22), advanced T-stage (T2/T3: HR =1.42, 95% CI: 1.21–1.66), and nodal involvement (N1: HR =1.35, 95% CI: 1.14–1.61) (all P<0.05).

Subsequent multivariate analysis confirmed older age (≥50 years: HR =1.99, 95% CI: 1.53–2.60), White race (HR =1.28, 95% CI: 1.04–1.57), unmarried status (HR =1.81, 95% CI: 1.55–2.12), advanced T-stage (T2/T3: HR =1.79, 95% CI: 1.50–2.13), and nodal involvement (N1: HR =1.51, 95% CI: 1.25–1.82) as independent risk factors (all P<0.001). Conversely, chemotherapy (HR =0.44, 95% CI: 0.37–0.52), ER positivity (HR =0.75, 95% CI: 0.61–0.93), non-infiltrating duct carcinoma (HR =0.54, 95% CI: 0.36–0.80), and single primary tumors (HR =0.41, 95% CI: 0.35–0.48) were significant protective factors. Complete regression results are detailed in Tables 2,3. The analysis incorporating patients with missing data using multiple imputation yielded results similar to the primary findings (Tables S1,S2).

For CSS, univariate cox regression in the training cohort (n=9,346) identified significant associations between survival outcomes and multiple variables: advanced age (≥50 years: HR =1.57, 95% CI: 1.12–2.21), higher tumor grade (III/IV: HR =1.83, 95% CI: 1.41–2.37), ER positive (HR =0.50, 95% CI: 0.39–0.64), PR positive (HR =0.54, 95% CI: 0.42–0.69), one primary only (HR =0.53, 95% CI: 0.41–0.69), unmarried status (HR =1.85, 95% CI: 1.46–2.36), advanced T-stage (T2/T3: HR =2.09, 95% CI: 1.64–2.66), and nodal involvement (N1: HR =2.10, 95% CI: 1.64–2.70). Subsequent multivariate analysis confirmed older age (≥50 years: HR =1.69, 95% CI: 1.20–2.40), higher grade (III/IV: HR =1.35, 95% CI: 1.03–1.77), unmarried status (HR =1.79, 95% CI: 1.41–2.28), advanced T-stage (T2/T3: HR =1.77, 95% CI: 1.38–2.29), and nodal involvement (N1: HR =1.81, 95% CI: 1.40–2.35) as independent risk factors. Conversely, ER positivity (HR =0.68, 95% CI: 0.50–0.94), non-infiltrating duct carcinoma (HR =0.31, 95% CI: 0.13–0.74), and single primary tumors (HR =0.52, 95% CI: 0.40–0.68) emerged as significant protective factors. Complete regression results are detailed in Tables 4,5. The analysis incorporating patients with missing data using multiple imputation yielded results similar to the primary findings (Tables S3,S4). Furthermore, results from the competing risk model were similar to those from the aforementioned analyses (Tables S5,S6).

Nomogram construction

Using multivariate analysis of the training cohort, we developed nomograms predicting OS and CSS (Figures 1,2). Each prognostic factor was assigned a score between 0 and 100, reflecting its impact on the model’s predictive accuracy. By aggregating these scores for each patient, we derived a total point value that facilitated the estimation of 3-, 5-, and 7-year OS and CSS probabilities. Importantly, higher total scores were associated with a worse prognosis for patients.

Figure 1 Nomogram for OS prediction in the patients. ER, estrogen receptor; N, node; 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; T, tumor.

Model validation

In the training cohort, the OS nomogram achieved a significantly higher C-index than the AJCC 7th staging system [0.72 (95% CI: 0.69–0.74) vs. 0.55 (0.52–0.57)]. Time-dependent ROC analysis yielded AUCs of 0.720 (3-year), 0.710 (5-year), and 0.709 (7-year) (Figure 3A). Validation confirmed these findings: C-index=0.71 (0.67–0.71) with corresponding 3-/5-/7-year AUCs of 0.724, 0.726 and 0.724 (Figure 3B). Calibration plots demonstrated excellent agreement between predicted and observed outcomes in both cohorts (training: Figure 4A-4C; validation: Figure 4D-4F). DCA revealed beneficial clinical utility of the nomogram in training (Figure 5A-5C) and validation sets (Figure 5D-5F).

Figure 3 ROC curves of the nomogram for 3-, 5-, and 7-year OS in training set (A) and test set (B). AUC, area under the curve; OS, overall survival; ROC, receiver operating characteristic.

Figure 4 Calibration plots of training set for 3- (A), 5- (B) and 7-year (C) OS; calibration plots of validation set for 3- (D), 5- (E) and 7-year (F) OS. OS, overall survival.

Figure 5 Decision curve analysis of nomogram for the 3-, 5-, and 7-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.

The CSS nomogram demonstrated robust validation. In the training cohort, its C-index reached 0.71 (95% CI: 0.68–0.74), exceeding the AJCC system’s 0.64 (95% CI: 0.60–0.67). Time-dependent ROC analysis yielded AUCs of 0.730 (3-year), 0.711 (5-year), and 0.710 (7-year) for CSS (Figure S2A). Validation cohort results confirmed this superiority: C-index=0.70 (95% CI: 0.65–0.75) with corresponding 3-/5-/7-year AUCs of 0.727, 0.715, and 0.705 (Figure S2B). Calibration plots showed high concordance between predicted and observed outcomes in both cohorts (training: Figure S3A-S3C; validation: Figure S3D-S3F). DCA revealed beneficial clinical utility of the nomogram in training (Figure S4A-S4C) and validation sets (Figure S4D-S4F).

Therapeutic efficacy across risk-stratified subgroups

Patients were stratified into three risk groups using the X-tile prediction model: low-risk (OS <289, CSS <219), intermediate-risk (289≤ OS ≤339, 219≤ CSS ≤266), and high-risk (OS ≥339, CSS ≥266) (Figure S5A,S5B). Kaplan-Meier analysis demonstrated significant survival differences between groups, with higher-risk patients exhibiting poorer OS and CSS compared to the improved outcomes observed in the low-risk group (Figure S5C,S5D). Following PSM, we evaluated differential chemotherapy treatment effects on OS and CSS in subgroup analyses.

In the low-risk groups, patients receiving adjuvant chemotherapy or NAC with CR showed improved OS. However, NAC with non-CR (NCR) conferred no significant OS benefit (Figure 6). Neither adjuvant chemotherapy nor NAC provided significant CSS benefits in this group (Figure S6). Furthermore, NAC showed no significant survival benefit compared to adjuvant chemotherapy.

Figure 6 Different chemotherapy strategies in low-risk groups for OS. The comparison between patients who received adjuvant chemotherapy and those who did not receive adjuvant chemotherapy (A). The comparison between patients who received NAC with CR (B) and those who did not receive NAC. The comparison between patients who received NAC with NCR (C) and those who did not receive NAC. The comparison between patients who received NAC with CR and NCR (D). The comparison between patients who received NAC with CR and those who received adjuvant chemotherapy (E). The comparison between patients who received NAC with NCR and those who received adjuvant chemotherapy (F). CR, complete response; NAC, neoadjuvant chemotherapy; NCR, non-complete response; OS, overall survival.

In the middle-risk groups, patients treated with adjuvant chemotherapy or NAC showed improved OS compared to those receiving no chemotherapy (Figure 7). Similarly, adjuvant chemotherapy and NAC with CR were associated with improved CSS versus no chemotherapy, whereas NAC with NCR showed no significant CSS benefit (Figure S7). NAC demonstrated no significant survival advantage over adjuvant chemotherapy.

Figure 7 Different chemotherapy strategies in middle-risk groups for OS. The comparison between patients who received adjuvant chemotherapy and those who did not receive adjuvant chemotherapy (A). The comparison between patients who received NAC with CR (B) and those who did not receive NAC. The comparison between patients who received NAC with NCR (C) and those who did not receive NAC. The comparison between patients who received NAC with CR and NCR (D). The comparison between patients who received NAC with CR and those who received adjuvant chemotherapy (E). The comparison between patients who received NAC with NCR and those who received adjuvant chemotherapy (F). CR, complete response; NAC, neoadjuvant chemotherapy; NCR, non-complete response; OS, overall survival.

In the high-risk groups, both adjuvant chemotherapy and NAC improved OS compared to no chemotherapy (Figure 8). Adjuvant chemotherapy and NAC with CR also improved CSS versus no chemotherapy; NAC with NCR did not (Figure S8). Notably, patients achieving CR with NAC showed superior CSS compared to those receiving adjuvant chemotherapy (Figure S8). Additionally, NAC with CR resulted in significantly better OS and CSS than NAC with NCR.

Figure 8 Different chemotherapy strategies in high-risk groups for OS. The comparison between patients who received adjuvant chemotherapy and those who did not receive adjuvant chemotherapy (A). The comparison between patients who received NAC with CR (B) and those who did not receive NAC. The comparison between patients who received NAC with NCR (C) and those who did not receive NAC. The comparison between patients who received NAC with CR and NCR (D). The comparison between patients who received NAC with CR and those who received adjuvant chemotherapy (E). The comparison between patients who received NAC with NCR and those who received adjuvant chemotherapy (F). CR, complete response; NAC, neoadjuvant chemotherapy; NCR, non-complete response; OS, overall survival.

Discussion

The management of HER2-positive BC has been transformed by targeted therapies, yet significant heterogeneity in outcomes persists even within T1–3N0–1 cohorts (14). While BCS with radiotherapy is established for this population, optimal systemic therapy strategies—particularly regarding chemotherapy sequencing and patient selection—remain debated (15,16). Our study addresses this gap by developing validated nomograms for OS and CSS and elucidating differential chemotherapy benefits across molecularly defined risk strata.

Our analysis identified advanced T-stage (T2/T3), nodal involvement (N1), multiple primary tumors, unmarried status, and older age (≥50 years) as independent predictors of reduced OS—aligning with established BC biology. Consistent with previous studies, we found that advanced age is a risk factor for both OS and CSS (17). Notably, tumor sequence (multiple primaries) emerged as a novel prognostic factor, associated with increased mortality risk. A possible reason is the greater tumor aggressiveness in this population (18). For CSS, consistent with previous studies, higher tumor grade (III/IV) and ER/PR negativity were additional risk factors, underscoring the aggressive biology of HR-negative HER2⁺ disease (19,20). The protective effect of chemotherapy and single-primary status highlights the interplay between tumor burden and treatment responsiveness. These findings extend prior HER2⁺ BC prognostic models by integrating clinicopathological and demographic variables often omitted in TNM-centric approaches.

However, some of our findings, such as those for race and histology, appear to contrast with conventional expectations and warrant further discussion. The efficacy of HER2-positive BC is now primarily driven by anti-HER2 agents, which may fundamentally alter the behavior and susceptibility of non-invasive ductal carcinoma (IDC) subtypes compared to their HER2-negative counterparts. In the era of potent HER2-targeted therapy, prognosis may be more strongly dictated by HER2 status than by histological subtype (21). Furthermore, the success of antibody-drug conjugates like trastuzumab deruxtecan underscores a paradigm shift toward targeted cytotoxic delivery, reducing reliance on histology-based chemosensitivity (22). Additionally, the tumor microenvironment and immune contexture significantly influence treatment response. Chemotherapy can modulate the TME and activate immune effector cells such as natural killer cells, potentially enhancing anti-tumor efficacy across histological subtypes (23). Thus, the observed protective effect of non-IDC may reflect a synergy between HER2 blockade and chemotherapy-induced immunomodulation—a mechanism that merits further investigation in subtype-specific settings. Regarding the association of White race with poorer OS, this finding contrasts with some historical trends but may reflect complex, non-biological factors not fully captured in the SEER database. These could include disparities in access to or adherence to the full course of modern, costly HER2-targeted therapies, differences in co-morbidity burdens, or variations in socioeconomic factors that influence long-term OS, which our study was not granular enough to delineate. Finally, the identified risk associated with unmarried status is consistent with a substantial body of oncologic literature, often attributed to poorer social support systems, which may impact everything from timely diagnosis and treatment adherence to overall physical and mental health during cancer survivorship. Using X-tile-derived cutoffs, we stratified patients into low-, intermediate-, and high-risk groups with distinct survival trajectories. This stratification revealed critical nuances in chemotherapy efficacy. Low-risk patients derived no CSS benefit from chemotherapy (adjuvant or NAC), while OS improvements were restricted to those achieving CR with NAC. This suggests overtreatment may occur in this subgroup, consistent with de-escalation trends in low-risk HER2⁺ BC (24). Intermediate-risk patients benefited from both adjuvant chemotherapy and NAC with CR for OS and CSS, but NAC without CR showed no survival advantage. This mirrors the predictive value of CR in HER2⁺ disease and emphasizes response-adapted therapy (25). High-risk patients exhibited the most pronounced benefit from systemic therapy: NAC with CR significantly improved CSS over adjuvant chemotherapy, while NAC without CR underperformed adjuvant regimens. This supports NAC’s dual role in downstaging tumors and identifying chemo-sensitive biology (26). Notably, NAC’s superiority was confined to CR achievers, aligning with the CREATE-X paradigm where residual disease warrants treatment escalation (27). Furthermore, the pivotal KATHERINE trial, with its latest update, has now demonstrated a significant OS benefit, further solidifying the role of trastuzumab emtansine (T-DM1) in the adjuvant setting. This phase III study confirmed that, compared to trastuzumab, adjuvant T-DM1 significantly improves OS in patients with HER2-positive early BC who have residual invasive disease after neoadjuvant therapy. This survival advantage builds upon the sustained improvement in invasive disease-free survival previously established, with T-DM1 reducing the risk of recurrence or death by 50% (28). Consequently, for the HER2-positive cohort in our study with high-risk features, the use of T-DM1 represents the standard-of-care for patients who do not achieve a pathologic complete response (pCR), directly addressing the unmet need for effective, escalated therapy in this population. This paradigm is distinct from the CREATE-X regimen, which is indicated for triple-negative BC, and underscores the critical importance of biomarker-specific treatment escalation. Our data suggest that high-risk patients with chemo-sensitive tumors (evidenced by CR) gain maximal benefit from NAC, whereas chemo-resistant tumors may fare better with adjuvant therapy—possibly due to eradication of micrometastases post-surgery (29).

Our OS and CSS nomograms demonstrated robust discrimination (C-index: 0.72 and 0.71, respectively), outperforming AJCC 7th edition staging (C-index: 0.55 and 0.64) and maintaining accuracy in validation cohorts. The models integrate 10 clinically accessible variables, enabling individualized risk estimation without genomic testing. High calibration accuracy and positive net benefit on DCA further support clinical adoption. For HER2⁺ T1–3N0–1 patients—a group where overtreatment and undertreatment coexist—these tools could guide decisions regarding chemotherapy sequencing, NAC candidacy, and post-NAC therapy intensification.

The differential efficacy of NAC according to pathological CR status likely reflects tumor-intrinsic biological factors: CR signifies chemosensitivity and HER2-dependency, whereas NCR tumors may harbor resistance pathways that impair NAC efficacy. Given the absence of NAC benefit in NCR patients within middle- and high-risk subgroups, exploration of biomarker-driven adjuvant strategies appears warranted. Several limitations merit consideration, including inherent constraints of the SEER database—notably the absence of HER2-targeted therapy data, which may confound chemotherapy effect estimates, limited granularity in NAC response assessment (such as residual cancer burden scores), and inability to capture recurrence dynamics or patient-reported outcomes like quality-of-life metrics. The absence of this treatment variable means that our cohort is inherently heterogeneous, comprising a mixture of patients who likely received the standard-of-care, suboptimal, or no targeted therapy. Instead, the primary value of this study should be viewed as a real-world analysis of population-level trends and outcomes during a pivotal transitional period in oncology. Future studies incorporating detailed treatment data are imperative to develop prognostic tools with genuine clinical utility in the modern targeted therapy era. The proposed model was developed and validated using only internal temporal validation. Therefore, its performance and applicability across different geographic regions or healthcare institutions remain unconfirmed and require further external validation. Despite propensity score matching, residual confounding from unmeasured variables—particularly socioeconomic factors influencing treatment access—remains possible. Validation in cohorts with documented HER2-targeted therapy use is critical. Integrating genomic classifiers could refine risk prediction. Additionally, evaluating NAC/adjuvant thresholds using nomogram scores—rather than rigid TNM criteria—may optimize therapy personalization. For patients with NCR, future trials exploring adjuvant strategies guided by nomogram risk strata are warranted. Specifically, investigations into antibody-drug conjugates (ADCs)—a class of targeted agents that deliver potent cytotoxic drugs directly to cancer cells via tumor-specific antibodies—are highly relevant. This approach aims to maximize efficacy while minimizing systemic toxicity, making it a promising strategy for overcoming treatment resistance in high-risk NCR patients.

In conclusion, we developed and validated prognostic nomograms that significantly outperform conventional staging in HER2-positive T1–3N0–1 BC. Our risk stratification identifies patients most likely to benefit from NAC (high-risk, CR-achievers) versus adjuvant chemotherapy (chemo-resistant subgroups). These tools address an unmet need for precision in early HER2⁺ BC management, in which treatment intensity must balance survival gains against toxicity risks. Prospective validation and integration of molecular biomarkers could further enhance their utility in guiding multidisciplinary therapy decisions.

Conclusions

Collectively, this study developed and validated prognostic nomograms that significantly outperform conventional AJCC 7th edition staging in HER2-positive T1–3N0–1 BC patients undergoing BCS. Our risk-stratification tool—integrating clinicopathological variables including tumor sequence, nodal status, and receptor profiles—identified distinct subgroups with differential therapeutic responses. Key findings demonstrate that while adjuvant chemotherapy provides broad survival benefits, patients with high-risk disease are more likely to derive superior survival from NAC. Conversely, low-risk patients, who derive limited benefit, may be candidates for chemotherapy de-escalation. These insights address a critical unmet need in HER2⁺ disease management: optimizing therapy intensity to balance survival gains against treatment-related toxicities. Our nomograms provide a clinically actionable framework for personalizing chemotherapy sequencing in this population, ultimately advancing precision oncology in HER2⁺ BC.

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-2025-302/rc

Peer Review File: Available at https://gs.amegroups.com/article/view/10.21037/gs-2025-302/prf

Funding: This research was funded by the National Natural Science Foundation of China (grant 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-2025-302/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. Gradishar WJ, Moran MS, Abraham J, et al. Breast Cancer, Version 3.2024, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw 2024;22:331-57. [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. Agostinetto E, Curigliano G, Piccart M. Emerging treatments in HER2-positive advanced breast cancer: Keep raising the bar. Cell Rep Med 2024;5:101575. [Crossref] [PubMed]
4. Li YW, Dai LJ, Wu XR, et al. Molecular Characterization and Classification of HER2-Positive Breast Cancer Inform Tailored Therapeutic Strategies. Cancer Res 2024;84:3669-83. [Crossref] [PubMed]
5. Veeraraghavan J, De Angelis C, Gutierrez C, et al. HER2-Positive Breast Cancer Treatment and Resistance. Adv Exp Med Biol 2025;1464:495-525. [Crossref] [PubMed]
6. Yan J, Xie Y, Liu Z, et al. DLL4-targeted CAR-T therapy sensitizes neoadjuvant chemotherapy via eliminating cancer stem cells and reshaping immune microenvironment in HER2(+) breast cancer. J Immunother Cancer 2024;12:e009636. [Crossref] [PubMed]
7. Kuemmel S, Graeser M, Schmid P, et al. Chemotherapy-free neoadjuvant pembrolizumab combined with trastuzumab and pertuzumab in HER2-enriched early breast cancer (WSG-KEYRICHED-1): a single-arm, phase 2 trial. Lancet Oncol 2025;26:629-40. [Crossref] [PubMed]
8. Boman C, Liu X, Eriksson Bergman L, et al. A population-based study on trajectories of HER2 status during neoadjuvant chemotherapy for early breast cancer and metastatic progression. Br J Cancer 2024;131:718-28. [Crossref] [PubMed]
9. Li JJ, Wang ZH, Chen L, et al. Efficacy and safety of neoadjuvant SHR-A1811 with or without pyrotinib in women with locally advanced or early HER2-positive breast cancer: a randomized, open-label, phase II trial. Ann Oncol 2025;36:651-9.
10. Bischoff H, Espié M, Petit T. Unveiling Neoadjuvant Therapy: Insights and Outlooks for HER2-Positive Early Breast Cancer. Curr Treat Options Oncol 2024;25:1225-37. [Crossref] [PubMed]
11. Stjepanovic N, Kumar S, Jerzak KJ, et al. Analysis of Factors Associated With Pathological Complete Response in Patients With HER2-Positive Breast Cancer Receiving Neoadjuvant Chemotherapy. Clin Breast Cancer 2024;24:e723-30. [Crossref] [PubMed]
12. Yee EK, Hallet J, Look Hong NJ, et al. Impact of Location of Residence and Distance to Cancer Centre on Medical Oncology Consultation and Neoadjuvant Chemotherapy for Triple-Negative and HER2-Positive Breast Cancer. Curr Oncol 2024;31:4728-45. [Crossref] [PubMed]
13. Teng L, Du J, Yan S, et al. A novel nomogram and survival analysis for different lymph node status in breast cancer based on the SEER database. Breast Cancer 2024;31:769-86. [Crossref] [PubMed]
14. Xu L, Xie Y, Gou Q, et al. HER2-targeted therapies for HER2-positive early-stage breast cancer: present and future. Front Pharmacol 2024;15:1446414. [Crossref] [PubMed]
15. Kwee E, de Groot LG, Alonso PR, et al. Neuropathic Pain Following Breast-conserving Surgery: A Systematic Review and Meta-Analysis. JPRAS Open 2024;42:48-57. [Crossref] [PubMed]
16. Metzger Filho O, Ballman K, Campbell J, et al. Adjuvant Dose-Dense Chemotherapy in Hormone Receptor-Positive Breast Cancer. J Clin Oncol 2025;43:1229-39. [Crossref] [PubMed]
17. Zheng Y, Yuan Y, Jin M, et al. Nomogram prediction of overall survival in breast cancer patients post-surgery: integrating SEER database and multi-center evidence from China. Front Oncol 2024;14:1470515. [Crossref] [PubMed]
18. Zhang BX, Brantley KD, Rosenberg SM, et al. Second primary non-breast cancers in young breast cancer survivors. Breast Cancer Res Treat 2024;207:587-97. [Crossref] [PubMed]
19. Yang R, Wu Y, Qi Y, et al. A nomogram for predicting breast cancer specific survival in elderly patients with breast cancer: a SEER population-based analysis. BMC Geriatr 2023;23:594. [Crossref] [PubMed]
20. Huang C, Ding Z, Li H, et al. A novel nomogram for predicting long-term heart-disease specific survival among older female primary breast cancer patients that underwent chemotherapy: A real-world data retrospective cohort study. Front Public Health 2022;10:964609. [Crossref] [PubMed]
21. Fan Z, Yuan Y, Chen X, et al. Anti-HER2 plus endocrine therapy versus anti-HER2 plus chemotherapy in hormone receptor-positive and HER2-positive metastatic breast cancer: a retrospective study. Int J Clin Oncol 2025;30:2532-40. [Crossref] [PubMed]
22. Modi S, Jacot W, Iwata H, et al. Trastuzumab deruxtecan in HER2-low metastatic breast cancer: long-term survival analysis of the randomized, phase 3 DESTINY-Breast04 trial. Nat Med 2025; Epub ahead of print. [Crossref]
23. Liu Z, Yang Z, Wu J, et al. A single-cell atlas reveals immune heterogeneity in anti-PD-1-treated non-small cell lung cancer. Cell 2025;188:3081-3096.e19. [Crossref] [PubMed]
24. Bolze A, Cirulli ET, Hajek C, et al. The Potential of Genetics in Identifying Women at Lower Risk of Breast Cancer. JAMA Oncol 2024;10:236-9. [Crossref] [PubMed]
25. Vasigh M, Karoobi M, Williams AD, et al. Neoadjuvant Endocrine Therapy Compared to Neoadjuvant Chemotherapy in Node-Positive HR+, HER2- Breast Cancer (Nodal pCR and the Rate of ALND): A Systematic Review and Meta-Analysis. Breast J 2024;2024:8866756. [Crossref] [PubMed]
26. Schettini F, Saracchini S, Bassini A, et al. Prediction of response to neoadjuvant chemotherapy by MammaTyper® across breast cancer subtypes: A retrospective cross-sectional study. Breast 2024;76:103753. [Crossref] [PubMed]
27. Lin K, Michaels E, Polley E, et al. Retrospective evaluation of adjuvant capecitabine dosing patterns in triple negative breast cancer. J Oncol Pharm Pract 2024; Epub ahead of print. [Crossref]
28. 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]
29. Cabioglu N, Karanlik H, Igci A, et al. Breast Cancer Recurrence in Initially Clinically Node-Positive Patients Undergoing Sentinel Lymph Node Biopsy After Neoadjuvant Chemotherapy in the NEOSENTITURK-Trials MF18-02/18-03. Ann Surg Oncol 2025;32:952-66. [Crossref] [PubMed]