Journal of Personalized Medicine (Dec 2022)

Web-Based Dynamic Nomograms for Predicting Overall Survival and Cancer-Specific Survival in Breast Cancer Patients with Lung Metastases

  • Kangtao Wang,
  • Yuqiang Li,
  • Dan Wang,
  • Zhongyi Zhou

DOI
https://doi.org/10.3390/jpm13010043
Journal volume & issue
Vol. 13, no. 1
p. 43

Abstract

Read online

Background: 60–70% of patients who die from breast cancer have lung metastases. However, there is a lack of readily available tools for accurate risk stratification in patients with breast cancer lung metastases (BCLM). Therefore, a web-based dynamic nomogram was developed for BCLM to quickly, accurately, and intuitively assess overall and cancer-specific survival rates. Methods: Patients diagnosed with BCLM between 2004 and 2016 were extracted from the Surveillance, Epidemiology, and Final Results (SEER) database. After excluding incomplete data, all patients were randomly assigned to training and validation cohorts (2:1). Patients’ basic clinical information, detailed pathological staging and treatment information, and sociological information were included in further analysis. Nomograms were constructed following the evaluations of the Cox regression model and verified using the concordance index (C-index), calibration curves, time-dependent receiver operating characteristic (ROC) curves, and decision curve analysis (DCA). Web-based dynamic nomograms were published online. Results: 3916 breast cancer patients with lung metastases were identified from the SEER database. Based on multivariate Cox regression analysis, overall survival (OS) and cancer-specific survival (CSS) are significantly correlated with 13 variables: age, marital status, race, grade, T stage, surgery, chemotherapy, bone metastatic, brain metastatic, liver metastatic, estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2). These are included in the construction of the nomogram of OS and CSS. The time-dependent receiver operating characteristic curve, decision curve analysis, consistency index, and calibration curve prove the distinct advantages of the nomogram. Conclusions: Our web-based dynamic nomogram effectively integrates patient molecular subtype and sociodemographic characteristics with clinical characteristics and guidance and can be easily used. ER-Negative should receive attention in diagnosing and treating BCLM.

Keywords