Background and Aims: Over the recent years, the incidence of breast cancer is increasingly shifting to younger population, which is more likely to develop axillary lymph node (ALN) metastasis. Therefore, this study was conducted to determine the influencing factors for ALN metastasis in young breast cancer patients using big-data platform of clinicopathologic information and establish a risk prediction models, so as to provide a reference for the clinical diagnosis and treatment of breast cancer in young adults.  
Methods: The clinicopathologic data of young patients who were diagnosed with breast cancer and underwent ALN dissection between 2010 and 2015 were selected from the SEER database. The influencing factors for ALN metastasis were determined by univariate and multivariate analysis, and were subsequently visualized by nomogram. The ability of the nomogram to identify patients with different ALN status was quantized using the AUC/C-index. The internal verification of the prediction performance of the nomogram was estimated by bootstrap method (1 000 replicates with a random seed of 12). Furthermore, the data of young patients with newly diagnosed breast cancer from 2015 to 2017 in Zhongnan Hospital of Wuhan University were collected for external validation of the original model. 
Results: A total of 23 778 young patients with breast cancer was recruited from the SEER database, 39.6% of whom had ALN metastasis. Univariate Logistic regression analysis showed that age, race, location of primary tumor, pathological grade, tumor size, and presence or absence of the chest wall or skin invasion as well as the status of ER, PR and HER-2 were significantly associated with ALN metastasis (all P<0.001). Multivariate Logistic regression analysis showed that age, race, and marital status, laterality, location of primary tumor and grade, tumor size, and presence or absence of the chest wall or skin invasion as well as the status of the ER, PR were independent influencing factors for ALN metastasis (all P<0.05), based on which, the risk prediction model was established. The calibration curve of internal validation indicated a good consistency between the predicted value calculated by the model and the real value (AUC/C-index=0.716). A total of 391 young patients with breast cancer were clinically enrolled as external validation dataset, and 49.9% of them were found to have ALN metastasis at initial diagnosis. The of external validation showed the good predictive ability of the model (AUC/C-index=0.798).
Conclusion: The risk prediction model developed using the SEER database for ALN metastasis in young patients with breast cancers has good predictive ability, and it can be used as a reference in clinical practice for estimating ALN metastasis of patients.