RESUMO
PURPOSE: Brain metastases (BM) occur in 15-35% of patients with metastatic breast cancer, conferring poor prognosis and impairing quality of life. Clinical scores have been developed to classify patients according to their prognosis. We aimed to check the utility of the Breast Graded Prognostic Assessment (B-GPA) and its modified version (mB-GPA) and compare them in routine clinical practice. METHODS: This is an ambispective study including all patients with breast cancer BM treated in a single cancer comprehensive center. We analyzed the overall survival (OS) from BM diagnosis until death. The Kaplan-Meier method and Cox proportional hazard regression model were used in the analyses. ROC curves were performed to compare both scores. RESULTS: We included 169 patients; median age was 50 years. HER2-positive and triple negative patients were 33.7% and 20.7%, respectively. At the last follow-up, 90% of the patients had died. Median OS was 12 months (95% confidence interval 8.0-16.0 months). OS was worse in patients with > 3 BM and in patients with triple negative subtype. CONCLUSIONS: In our series, we confirm that B-GPA and mB-GPA scores correlated with prognosis. ROC curves showed that B-GPA and mB-GPA have similar prognostic capabilities, slightly in favor of mB-GPA.