Development, verification, and comparison of a risk stratification model to identify potential population benefiting from chemotherapy in non-metastatic male breast cancer

Abstract

To establish a prognostic stratification nomogram for non-metastatic male breast cancer to determine which patients can benefit from chemotherapy. A population-based study was conducted using data collected from the surveillance, epidemiology, and end results database. Cox proportional hazards analysis identified significant prognostic factors for survival. A prognostic stratification model was developed using R software. Propensity score matching was implemented to balance characteristics between the chemotherapy cohort and the non-chemotherapy cohort. The multivariate analyses indicated that age, race, grade, surgery, primary tumor, marital status, T stage, and N stage were independent prognostic factors for overall survival in non-metastatic metastatic MBC patients who did not receive chemotherapy (all P < 0.05). The C-index was 0.786 (95% CI 0.662-0.870) in the training cohort and 0.763 (95% CI 0.517-0.852) in the validation cohort. The nomogram effectively discriminated between low-risk, moderate-risk, and high-risk groups concerning OS (P < 0.0001). The current study developed the first prognostic stratification nomogram for non-metastatic MBC and identified that patients in the moderate-risk and high-risk groups are more likely to benefit from chemotherapy.

Keywords: Chemotherapy; Male breast cancer; Nomogram; Overall survival; Prognosis; SEER.

Fig. 1

Flowchart of the study design. A total of 4445 patients with non-metastatic male breast cancer were finally involved in our study.

Fig. 2

Development of a prognostic stratification nomogram and validation of the proposed nomogram. (A) A prognostic stratification nomogram to accurately predict overall survival for non-metastatic MBC. (B) Identify the optimal cutoff value for age using R software.

Fig. 3

Validation of the proposed nomogram. (A,B) ROC curves for predicting the overall survival in the training and validation cohorts at 3-year, 5-year, and 10-year, respectively. (C,D) the time-dependent AUC values of nomogram in the training cohort and validation cohort. (E,F) the calibration curves for predicting patients’ overall survival in the training and validation cohorts at 3-year, 5-year, and 10-year, respectively.

Fig. 4

Decision curve analysis for non-metastatic MBC. Decision curve analysis of the nomogram and single independent predictors for predicting the 3-year OS (A,D), 5-year OS (B,E) and 10-year OS (C,F) in the training cohort and validation cohort, respectively.

Fig. 5

Determination of best cut-off value of total sum score by the X-tile software (A–C). The cut-point distribution histogram illustrates the distribution of patient numbers corresponding to each threshold value across the range of cut-off points (A). The cut-point selection plot, generated by X-tile, visually presents the distribution of low-population and high-population patient groups across various threshold values, assisting researchers in identifying an optimal cutoff that is both statistically significant and clinically meaningful (B). The survival curve plot illustrates the differences in survival between patient groups stratified by a specific cutoff value, aiding in the evaluation of the clinical and statistical significance of the selected threshold (C). Kaplan–Meier survival curves between the chemotherapy cohort and non-chemotherapy cohort for the entire group before PSM (D) and after PSM (F). Survival curves in the entire cohort stratified by the total score of the nomogram (E).

Fig. 6

Kaplan–Meier survival curves of OS between the chemotherapy cohort and non-chemotherapy cohort for the low-risk group (A,D), moderate-risk group (B,E) and high-risk group (C,F) before PSM and after PSM.