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

Gland Surg. 2025 Nov 30;14(11):2302-2320. doi: 10.21037/gs-2025-302. Epub 2025 Nov 25.

ABSTRACT

BACKGROUND: Despite advances in human epidermal growth factor receptor 2 (HER2)-targeted therapies, HER2-positive early-stage breast cancer (BC) exhibits substantial outcome heterogeneity after breast-conserving surgery (BCS), complicating personalized chemotherapy sequencing. Robust tools integrating clinicopathological variables and treatment response are urgently needed to optimize survival while minimizing overtreatment. This study aimed to develop and validate a novel prognostic tool for this specific patient population.

METHODS: Using Surveillance, Epidemiology, and End Results (SEER) data [2010-2016; HER2-positive T1-3N0-1 BC patients (n=13,875) treated with BCS and radiotherapy], we developed and validated nomograms predicting overall and cancer-specific survival (CSS). Multivariable Cox regression identified prognostic factors, with nomogram performance evaluated via concordance index (C-index), time-dependent area under the curve (AUC), calibration, and decision curve analysis (DCA). Risk stratification used X-tile-derived thresholds.

RESULTS: Key independent predictors included tumor sequence [multiple primaries: overall survival (OS) hazard ratio (HR) =2.70], T stage (T2/T3: OS HR =1.79), nodal involvement (N1: OS HR =1.51), marital status (unmarried: OS HR =1.81), and chemotherapy administration (protective OS HR =0.44). Nomograms significantly outperformed American Joint Committee on Cancer (AJCC) 7th staging (OS C-index: 0.72 vs. 0.55; CSS C-index: 0.71 vs. 0.64; both P<0.001). Patients were stratified into low-, intermediate-, and high-risk groups. Adjuvant chemotherapy demonstrated improved OS in low- and middle-risk groups compared with no chemotherapy, though it showed no significant benefit for CSS. Similarly, neoadjuvant chemotherapy (NAC) improved OS only in low- and middle-risk group patients who achieved a complete response (CR), but likewise conferred no significant CSS benefit. High-risk patients achieving NAC-induced CR showed superior CSS versus adjuvant chemotherapy, while non-CR (NCR) patients derived no survival benefit. Within the high-risk group, patients who achieved CR demonstrated significantly improved OS and CSS compared to those with NCR.

CONCLUSIONS: We present the first validated nomograms integrating tumor multiplicity, sociodemographic factors, and treatment response for HER2-positive BC. "Multiple primary tumors" emerged as a novel prognostic indicator, suggesting unexplored biological aggression. Our risk stratification translates to actionable strategies: chemotherapy de-escalation may be warranted in low-risk patients, as they derive no clear CSS benefit from NAC. Conversely, patients within the high-risk category are more likely to benefit from NAC and should be prioritized for this treatment to maximize survival gains. Prospective integration of targeted therapy data will refine these precision oncology tools.

PMID:41377890 | PMC:PMC12685774 | DOI:10.21037/gs-2025-302