Abstract Uterine serous carcinoma (USC) is a rare but particularly invasive histological subtype of uterine cancer, the most common gynecologic malignancy in developed countries. Though previous studies with small sample sizes have been conducted to identify transcriptomic and proteomic biomarkers for USC, none have resulted in a clinical assay for patient risk stratification. I have discovered a panel of 78 genes highly upregulated in poor prognosis patients from The Cancer Genome Atlas (TCGA) USC dataset. To create a composite score for our biomarker panel, I applied a machine learning algorithm (elastic net regression) to the TCGA USC gene expression data set to generate a model which outputs a linear predictor score (USC78) based on individual patients’ expression of the 78 genes from our signature. I then quantified gene expression from formalin-fixed paraffin embedded (FFPE) tumor tissue in our cohort of Augusta University (AU) patients and calculated their USC78 score to demonstrate the ability of this model to separate good and poor survival prognosis in a single-center, retrospective validation study. In both TCGA and AU, higher USC78 scores are associated with worse overall survival. This score is also able to risk stratify serous ovarian carcinoma patients in the TCGA data set, suggesting that USC78 may be an important prognosis prediction tool in serous gynecological cancers. Additionally, network analysis of these genes reveals increased TGF-B signaling correlates with the poor-prognosis USC78-high tumor expression profile. TGF-B inhibition has demonstrated chemosensitization effects in primary USC tumor cell lines. I aim to use USC78 to better understand the biological mechanisms behind poor survival prognosis in USC patients and identify potential therapeutic targets for USC. One such target is TGF-B inhibition, and further knockdown, overexpression, and inhibitor studies will be conducted to test whether TGF-B inhibition can improve USC patient prognosis. Citation Format: Lynn K. Tran, Emily K. Myers, David P. Mysona, Paul M. Tran, Wonsok Lee, John J. Wallbillich, Daniel Kleven, Sharad Ghamande, Jin-Xiong She. Predicting survival and improving treatment for uterine serous carcinoma patients using USC78, a genomic risk score [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1704.