Continued Endocrine Therapy Is Associated with Improved Survival in Patients with Breast Cancer Brain Metastases.

Abstract

Brain metastases (BMs) are a rare but devastating condition in estrogen receptor (ER)-positive metastatic breast cancer (MBC). Although endocrine therapy (ET) is the mainstay of treatment in this disease subtype, only case reports have been published concerning the activity of ET in BMs henceforth. Therefore, we aimed to systematically investigate the impact of ET after diagnosis of BM on outcome and clinical course of disease in patients with ER-positive MBC. Patient characteristics, detailed information about BMs including diagnosis-specific graded prognostic assessment class (DS-GPA), and clinical outcome were obtained by retrospective chart review for all patients treated for ER-positive breast cancer BMs between 1990 and 2017 at an academic care center. Overall survival (OS) was measured as the interval from diagnosis of BM until death or last date of follow-up. Overall, 198 patients [female: 195/198 (98.5%); male: 3/198 (1.5%)] with ER-positive breast cancer BMs were available for this analysis. Eighty-eight of 198 patients (44.4%) received ET after diagnosis of BM including aromatase inhibitors (AIs; letrozole, anastrozole, exemestane), tamoxifen, and fulvestrant. Median OS was significantly longer in patients receiving ET after diagnosis of BM compared with patients who did not (15 vs. 4 months, <i>P</i> < 0.001; log-rank test). No significant difference in terms of OS was observed between patients receiving AIs, tamoxifen, or fulvestrant. In patients with concomitant leptomeningeal carcinomatosis (LC), ET prolonged median OS significantly as well (7 vs. 3 months, <i>P</i> = 0.012; log-rank test). In a multivariate analysis including DS-GPA and ET, only treatment with ET after diagnosis of BM (HR, 0.69; 95% confidence interval, 0.48-0.99; <i>P</i> = 0.046) was associated with prognosis (Cox regression model). Continuing ET after BM diagnosis was associated with a significantly prolonged OS in this large single-center cohort. No substantial differences between substances were observed. These findings should be validated in a prospective cohort.