Prognostic Factors Analysis in Breast Cancer Patients with Brain Metastases: Identification of Key Determinants of Survival

Background : With advances in systemic therapy, survival has improved in patients with breast cancer brain metastases (BCBM). Radiotherapy (RT), as a local treatment, further contributes to this improvement. This study investigates the prognostic factors for survival in BCBM patients with a Karnofsky Performance Status (KPS) score ≥70 who underwent craniocerebral radiotherapy (CRT). Methods : A total of 505 patients with BCBM diagnosed in our hospital from January 2017 to December 2023 were retrospectively explored. Subsequently, Kaplan-Meier method (K-M method) was used to analyze overall survival (OS) and progression-free survival (PFS). Additionally, Clinical data were collected, radiomics and dosimetric factors were extracted for univariate and multivariate model analysis, and a survival prediction model was established. Results : The median overall survival (mOS) of the 505 BCBM patients was 20.8 months (95% CI 18.8 - 23.2 months). Multivariate analysis identified the number (≤4) of brain metastases (P<0.001) and CRT (P<0.001) as independent prognostic factors. Specifically, the median OS in the CRT group (n=293) was 27.0 months (95% CI 23.33 - 30.50 months), which was significantly longer than that of the non-radiotherapy group (n=201), which had a median OS of 15.10 months (95% CI 10.97 - 17.17 months). A significant survival difference was observed between the two groups (P < 0.001, HR 0.522, 95% CI 0.424-0.632). Furthermore, Multivariate analysis revealed that leptomeningeal metastasis (LM) (P=0.015), extracranial metastasis (P=0.015), and intracranial progression after RT (P=0.013) were independent adverse prognostic factors for OS in patients receiving CRT. Additionally, the median progression-free survival (PFS) in the radiotherapy group was 16.3 months (95% CI 14.6 - 19.17 months), with LM (P<0.001) identified as an independent adverse prognostic factor for PFS. In the subgroup analysis, patients with positive expression of human epidermal growth factor receptor 2 (HER2) had a better survival benefit when treated with targeted therapy (TT) (P=0.021). Finally, a predictive model incorporating radiomic and radiodosimetric features (n=149) was developed and showed excellent performance in predicting 1-, 2-, and 3-year survival, with area under curve (AUC) values of 0.965, 0.861, and 0.859, respectively. Conclusion : For BCBM patients with a KPS ≥70, CRT is associated with a significant survival benefit. However, the development of LM remains a major adverse factor leading to poorer overall prognosis. Besides, the integration of radiomic and radiodosimetric features into predictive models demonstrates strong potential for accurately estimating prognosis. These findings suggest that such models could enhance clinical decision-making, ultimately supporting improved patient outcomes.