Background: Racial disparities in breast cancer (BC) mortality are attributed to later stage diagnoses and a higher incidence of triple-negative BC among African American (AA) women. In previous work, we showed that AA women with ER+ BC are more likely to develop biologically aggressive disease and are more likely to die from early stage, ER+ BC than non-Hispanic White women (Hoskins et al, JAMA Oncol, 2021). The underlying molecular drivers of this disparity are unknown. Here we report the molecular characterization of a series of luminal BC from AA women. Methods: Consecutive breast tumor specimens received in the Pathology Department underwent next generation sequencing (NGS). Unstained FFPE tissue sections were macro-dissected to isolate tumor cells, and nucleic acids were extracted using commercially available kits. DNA and RNA sequencing libraries were prepared with the Oncomine Comprehensive Assay v3 (OCAv3) (Thermo Fisher), which includes 161 driver genes and detects SNVs, CNVs, INDELs and gene fusions. Sequencing was performed on the Ion S5XL sequencer. Sequencing reads were mapped to the UCSC human genome build GRCh37/hg19 using Torrent Suite™ software (version 5.10; Thermo Fisher). Data analysis and variant calling was performed using the Ion Reporter analysis tool. Results: We identified 60 somatic driver gene alterations in luminal tumors from 35 AA patients (primary tumors, n = 26; metastatic tumors, n = 9). Recurrently altered genes identified in > 5% of tumors are listed in the Table. The most frequently altered gene was PIK3CA (42% of tumors). ESR1 gene fusions were seen in 25% of tumors. Interestingly, an equal frequency of ESR1 fusions were detected in primary (27%) and metastatic (22%) tumors, in contrast to activating mutations which are found in recurrent tumors following treatment with aromatase inhibitors. ARID1A alterations were identified in 17% of primary tumors. ARID1A encodes a subunit of the SWI/SNF chromatin remodeling complex. Alterations in ARID1A confer endocrine resistance, and are enriched in recurrent tumors in the literature. We also found a high number of CNVs in members of the FGF gene family (36% of tumors), which are also associated with resistance to endocrine therapy. An in silico analysis comparing our findings with publicly available datasets will be presented. Conclusions: This study of somatic driver gene alterations in a consecutive series of luminal breast tumors from AA patients found a higher than expected frequency of alterations in genes associated with endocrine resistance in untreated primary tumors, suggesting a partial explanation for racial disparities in survival.