Effect of prior breast cancer on survival of female patients with primary liver cancer: Development of a competing risk model nomogram

Abstract

Abstract Background The impact of prior breast cancer on subsequent primary liver cancer (PLC) survival remains poorly understood. Moreover, traditional prediction models struggle to accurately predict cancer-specific survival (CSS) for PLC cases that have a history of breast cancer. We aimed to investigate the role of prior breast cancer on subsequent PLC survival and construct a CSS prediction nomogram for PLC cases with a history of breast cancer. Methods We obtained data on female PLC patients between 2005 and 2015 from the Surveillance, Epidemiology, and End Results (SEER) database. To minimize the impact of confounding bias, we employed propensity score matching (PSM) to match each patient with prior breast cancer to 10 patients without a history of breast cancer. Univariate, as well as multivariate COX survival and CSS analyses, were conducted to investigate the effect of prior breast cancer on subsequent PLC survival. Additionally, a competing risk model nomogram was built to predict PLC-specific survival. Results Our survival analyses revealed that prior breast cancer did not significantly affect overall survival (OS) among PLC cases. However, it served as a prognostic factor for predicting favorable outcomes in PLC-specific survival. A history of prior breast cancer reduced PLC-specific mortality by 0.26-fold (HR = 0.74, 95% CI: 0.88–0.96, p = 0.023). Furthermore, the analysis of concordance index (C-index), receiver operating characteristic (ROC) curves and calibration curves showed that our model had good predictive power and outperformed conventional prediction models. According to decision curve analysis (DCA), our constructed nomogram had good clinical significance. Conclusions Prior breast cancer is beneficial to PLC-specific survival in PLC patients. The constructed competing risk model nomogram demonstrated good predictive ability for PLC-specific survival.