Background For primary neuroendocrine carcinoma of breast was a very rare subtype in breast cancers, its prognosis was still controversial and there was no independent standard for its treatment. The purpose of our retrospective study was to construct a nomogram to predict the overall survival (OS) of patients with neuroendocrine carcinoma of the breast.
Methods 150 patients of training cohort were collected from Surveillance, Epidemiology, and End Results (SEER) database diagnosed between 2003 and 2015, and 93 patients of verification cohort were enrolled from Sun Yat-sen University Cancer Center (Guangzhou, China) diagnosed between 2004 and 2018. The nomogram was constructed uniting three significantly risk factors of overall survival identified by univariate and multivariate analysis and then validated using receiver operating characteristic (ROC) curves for discrimination, calibration plots and the decision curves analysis (DCA).
Results Age, N stage and PR status were closely and significantly related to overall survival in patients with breast neuroendocrine carcinoma. The C-index of nomogram in the training and verification cohorts are 0.775 (95% CI, 0.784 to 0.615) and 0.760 (95% CI, 0.705 to 0.800) respectively. Calibration plots of practical and predicted possibility for the nomogram demonstrated that the predictive 5-year overall survival rate was in accordance with the actual overall survival probability in both sets. Moreover, the decision curves (DCA) also expressed pretty clinical benefit of the nomogram across a range of high-risk threshold.
Conclusion This novel population-based nomogram may help with treatment decisions in patients with neuroendocrine carcinoma of the breast (NEBC).