Abstract Background: Metastatic breast cancer (MBC) is incurable, yet some extreme long-term survivors (ELTS) live for a decade or more. Mechanisms underlying extreme survival are incompletely known, yet are important to improve prognosis and treatment. Extreme survival also provides an opportunity for cancer cells to evolve over decades. To evaluate the genomic characteristics of ELTS, we identified a cohort representative of the extreme tail of the metastatic survival curve and performed genomic analysis of blood and matched metastatic tumor. Methods: We identified a cohort of ELTS at our institution in an IRB-approved study. Eligible subjects had MBC and had lived a ≥10 years from the original breast cancer (BC) diagnosis with hormone receptor (HR)-positive disease or ≥5 years if HR-negative. Those with available archived tumor tissue were enrolled for genomic analyses. Archived samples were obtained from metastatic sites obtained late in the clinical course. Whole exome sequencing (WES) was performed on matched blood and tumor. Mutational profiles were qualitatively compared with publicly available reference somatic alterations in primary BC (Ctrl; cbioportal.org). Frequency of mutations is reported for known BC drivers. In an exploratory analysis to identify novel genes, the frequency of mutations in ELTS vs. Ctrl was compared for 15,659 well-covered genes. These were further evaluated for the ratio of non-synonymous to synonymous mutations, which is indicative of selective pressure. Results: 53 ELTS were identified; 13 had available archived tumor tissue and consented. The median survival from original BC diagnosis was 22 years, which corresponds to 99.8th percentile for survival, benchmarked to a cohort of MBC patients identified from the Surveillance, Epidemiology, and End Results-Medicare linked database. The 13 subjects had lived with metastatic disease for a median of 9 years (range 4-23). One subject had HR-/HER2+ BC and had lived 16 years with MBC. Three additional subjects had HR+/HER2+ BC with survival ranging from 19-25 years from original BC diagnosis. The remaining 9 subjects had HR+/HER2- BC with survival of 13-40 years from original diagnosis. Matched tumor/somatic WES achieved a mean tumor read depth of ˜600x and 96% coverage. Somatic mutations revealed near-expected rates of known driver alterations in commonly mutated genes such as PIK3CA, TP53, and CDH1 (Table). No mutations were found in PTEN or ESR1 in ELTS. Novel gene candidates were identified as having a high prevalence of mutations in this cohort including Androgen Receptor (10 of 13), MUC4 (12 of 13) MUC12 (10 of 13), and HRNR (12/13). Pathway analysis is ongoing. Conclusions: ETLS tumors obtained late in the clinical course have mutations in known cancer drivers near rates expected from primary BC. However, we have identified additional genes with an increased number of mutations either related to distinct genetic features of these cancers, or due to accumulated mutations that occur over the extended time for tumor evolution. Frequency of driver mutations in control and extreme long-long term survivorsGeneCtrl freqELTS freqGeneCtrl freqELTS freqPIK3CA0.350.38MAP2K40.020.38TP530.350.23NF10.040.15GATA30.120.08CBFB0.030.15CDH10.100.23ERBB20.030.08 Citation Format: Yan RE, Longhurst C, Lee K, Gilbert AD, Alexandridis RA, McGregor SM, Laffin JJ, Rocque G, Burkard ME. Whole exome analysis of extreme long-term survivors with metastatic breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-06-14.