A nomogram for predicting survival in patients with de novo metastatic breast cancer: a population-based study

BACKGROUND: 5–10% of patients are diagnosed with metastatic breast cancer (MBC) at the initial diagnosis. This study aimed to develop a nomogram to predict the overall survival (OS) of these patients. METHODS: de novo MBC patients diagnosed in 2010–2016 were identified from the Surveillance, Epidemiology, and End Results (SEER) database. They were randomly divided into a training and a validation cohort with a ratio of 2:1. The best subsets of covariates were identified to develop a nomogram predicting OS based on the smallest Akaike Information Criterion (AIC) value in the multivariate Cox models. The discrimination and calibration of the nomogram were evaluated using the Concordance index, the area under the time-dependent receiver operating characteristic curve (AUC) and calibration curves. RESULTS: In this study, we included 7986 patients with de novo MBC. The median follow-up time was 36 months (range: 0–83 months). Five thousand three-hundred twenty four patients were allocated into the training cohort while 2662 were allocated into the validation cohort. In the training cohort, age at diagnosis, race, marital status, differentiation grade, subtype, T stage, bone metastasis, brain metastasis, liver metastasis, lung metastasis, surgery and chemotherapy were selected to create the nomogram estimating the 1-, 3- and 5- year OS based on the smallest AIC value in the multivariate Cox models. The nomogram achieved a Concordance index of 0.723 (95% CI, 0.713–0.733) in the training cohort and 0.719 (95% CI, 0.705–0.734) in the validation cohort. AUC values of the nomogram indicated good specificity and sensitivity in the training and validation cohort. Calibration curves showed a favorable consistency between the predicted and actual survival probabilities. CONCLUSION: The developed nomogram reliably predicted OS in patients with de novo MBC and presented a favorable discrimination ability. While further validation is needed, this may be a useful tool in clinical practice.