1039 Background: GATA3 expression is broadly used as a biomarker for diagnosing cancer of breast origin and its expression is strongly associated with estrogen receptor (ER)-positive luminal phenotype. Although GATA3 mutations are observed in 12-18% of breast cancers (BC), they are poorly characterized. Recently, the pharmacological inhibition of MDM2 has been shown to significantly impair tumor growth in GATA3-deficient models in vitro, in vivo, and patient-derived xenograft harboring GATA3 somatic mutations. Therefore, given the potential targetability of GATA3-mutated BC cells, it is important to better understand and characterize the mutational landscape of GATA3 to help guide future prospective clinical studies. Methods: We accessed a cohort of BC samples profiled genomically and transcriptomically from two open-source datasets: TCGA (n=961) and METABRIC (n=1866). We used the chi-squared test to analyze the frequency of GATA3 mutations across PAM50 subtypes (Basal, HER2, Luminal A, Luminal B, Normal-like) and histological BC subtypes (invasive ductal carcinoma [IDC] and invasive lobular carcinoma [ILC]). Mutations affecting other genes in the GATA3-mutated samples and the GATA3 wild-type (WT) samples were compared using the chi-squared test. We analyzed the mutational landscape of the GATA3 gene and correlated different sites of mutations with GATA3 expression, MDM2 amplification, other co-occurring mutations, and clinical behavior. GATA3 expression levels were compared using the Mann-Whitney test. Results: Of the analyzed 2,827 BC samples, GATA3 mutations were unevenly distributed in the five PAM50 subtypes, as Luminal A and B subtypes had the highest frequency (16% and 17%, respectively) and the basal subtype had no observed GATA3 mutations (p <0.0001). Further analysis showed that GATA3 mutations were more common in luminal IDC than ILC (17% vs 8.7%; p <0.0001). While mutations in CBFB, AKT1, TBX3 and ARID1A were more frequently co-occurring with GATA3 mutations, mutations in TP53, CDH1 and PIK3CA were mutually exclusive with GATA3 mutations (adjusted p <0.0001 for all genes). Analysis of the GATA3 mutational landscape identified two types of mutations in GATA3: a group of truncating mutations occurring at the end of the GATA3 gene (after the 307 position including the splice site at 308) and a group occurring before the 307 position. The first group was associated with increased GATA3 expression, MDM2 amplification, and was mutually exclusive with TP53 mutations. The second group had little to no effect on GATA3 expression and behaved similarly to GATA3 WT. Conclusions: A subtype of GATA3 mutations are mutually exclusive with TP53 mutations and are associated with increased MDM2 amplification, making them an ideal target for clinical trials involving MDM2 inhibitors.