Prognostic nomogram for female patients suffering from non-metastatic Her2 positive breast cancer: A SEER-based study

This paper aimed at constructing and validating a novel prognostic nomogram, so that physicians forecast the overall survival (OS) rates of female patients suffering from non-metastatic human epidermal growth element receptor-2 (HER2) positive breast. Information of primary female her2 positive breast cancer patients without metastasis was obtained from the Surveillance, Epidemiology, and End Results (SEER) database with given inclusion and exclusion standards. Independent variables were obtained greatly by performing univariable and multivariate analyses. Based on those independent predictors, a novel prognostic nomogram was constructed for predicting the survival of those with 3- and 5-year OS. Then, concordance index (C-index), receiver operating characteristic curve (ROC), and calibration plot were adopted for the assessment of the predictive power of the nomogram. A total of 36,083 eligible patients were classified into a training cohort (n = 25,259) and a verification cohort (n = 10,824) randomly. According to the identification of multivariate analysis, survival was predicted by age at diagnosis, marital status, race, site, T stage, N stage, progesterone receptor (PR) status, estrogen receptor (ER) status, surgery, radiation, and chemotherapy independently. A nomogram was established by applying the training cohort. The nomogram displayed excellent discrimination and performance as indicated by the C-index (0.764, 95% confidence interval: 0.756-0.772), and the 3- and 5-year area under the curve of ROC (AUC) values (0.760 and 0.692 respectively). The calibration plots for forecasting the 3- and 5-year OS were in great agreement. The OS for female her2 positive breast cancer patients without metastasis was predicted by constructing a nomogram on basis of the SEER database. A precise survival prediction could be offered for each patient.