Clinical and molecular features of resected breast cancer brain metastases: A retrospective cohort study.

e14013 Background: Brain metastases (BMs) are a significant cause of morbidity and mortality in patients (pts) with metastatic breast cancer (MBC). Some pts undergo surgical BM resection; however, clinical and molecular characteristics and prognosis in this group are poorly characterized. Methods: 107 pts with MBC who underwent surgical BM resection from 2006-2024 at a single institution were retrospectively identified. Review of electronic medical records identified baseline characteristics, treatment history, BM receptor status, next generation sequencing (NGS), and outcomes. Median overall survival (mOS) was estimated by the Kaplan-Meier method; cox proportional hazards model was used to evaluate factors associated with mOS. Results: Most pts (n = 106, 99.1%) were female with median age 56 years (range 26-84 years). Receptor subtypes at MBC diagnosis included hormone receptor (HR)+/HER2- (n = 27, 25.2%), HER2+ (n = 44, 41.1%), and triple negative BC (TNBC; n = 36, 33.6%). BMs were present at initial MBC diagnosis in most pts (n = 65, 60.8%). Median time from MBC diagnosis to index BM resection was longer for pts with HER2+ MBC (16.6 months) vs. TNBC and HR+/HER2- MBC (0.03 and 0.3 months, respectively, p = 0.01). At time of first BM resection, most pts had a single BM (n = 63, 58.9%) and half (n = 52, 48.7%) had extracranial disease. Most pts received postoperative radiation (n = 91, 85.1%) and systemic therapy (n = 85, 79.4%). Resected BM receptor status was available for 91 pts (85.0%); 27 pts (25.2%) had receptor discordance, mostly loss of HR positivity in the BM (n = 20, 74.1%). BM NGS testing was available for 40 pts (37.4%); most common pathogenic alterations were TP53 (n = 28, 70.0%), ERBB2 amplification (n = 15, 37.5%), and PIK3CA (n = 9, 22.5%). In 5 pts with both BM and peripheral NGS, targetable mutations were concordant in 2 pts and discordant in 3 pts. mOS from time of BM resection was shortest for pts with TNBC (14.4 months) vs. HR+/HER2- and HER2+ MBC (37 and 35 months respectively, p = 0.02). In a multivariate analysis, factors associated with longer mOS post-BM resection included receipt of postoperative stereotactic radiosurgery (HR 0.36, CI 0.15-0.86, p = 0.02) and postoperative systemic treatment (HR 0.18, CI 0.08-0.38, p < 0.01). Factors associated with shorter mOS included presence of systemic extracranial disease at time of resection (HR 2.97, CI 1.71-5.16, p < 0.01) and development of leptomeningeal disease (LMD) (HR 2.47, CI 1.28-4.75, p = 0.01). Conclusions: In a cohort of > 100 pts with MBC and resected BMs, pts with TNBC survived ~1 year post resection and pts with HR+/HER2- and HER2+ MBC survived ~3 years post resection. Extracranial disease at time of resection and development of LMD were associated with worse prognosis. In resected BMs, loss of HR positivity was common (~25%) and targetable mutations were frequent (~75%), supporting the utility of molecular analysis of resected BMs to guide management.