These challenges are significant, as they prevented us from defin- itively classifying these cancers as hereditary or sporadic. These include pretesting factors (n = 10, 4.7%; patient does not want testing; insurance denies reimbursement for testing; tumor insuffi- cient for testing), IHC/MSI discordances (n = 6, 2.8%), MSI-L with intact MMR protein expression by IHC (n = 5, 2.3%), and post-tissue testing issues (n = 8, 3.8%; insurance denies reimbursement for germline testing; patient declines germline testing; VUS detected; no germline mutation identified). Conclusion: Challenges associated with universal screening occurred in 13.6% of screened patients (29/213) that preclude designation of a tumor as sporadic or hereditary. doi:10.1016/j.ygyno.2017.03.190 163 - Poster Session CTNNB1 predicts better progression-free survival in endometrioid endometrial cancer patients treated with chemotherapy E.R. Penick a , b , c , , C. Tian b , N.W. Bateman c,d,e , J. Oliver b , D. Mitchell b , T.P. Conrads b,c,e,f,g , C.A. Hamilton a,b,c,e , G.L. Maxwell b,c,e,f,g , K.M. Darcy b,c,e . a Walter Reed National Military Medical Center, Bethesda, MD, USA, b Gynecologic Cancer Center of Excellence, Bethesda, MD, USA, c Uniformed Services University of the Health Sciences, Bethesda, MD, USA, d Gynecologic Cancer Center of Excellence, Annandale, VA, USA, e Murtha Cancer Center, Bethesda, MD, USA, f Inova Schar Cancer Institute, Fairfax, VA, USA, g Inova Fairfax Hospital, Falls Church, VA, USA Objective: CTNNB1 encodes β-catenin, a member of the Wnt signaling pathway. Mutations and overexpression of CTNNB1 are common in colon, ovarian, and endometrioid endometrial cancers (EEC). This study determined whether the relationship between endometrial cancer biomarkers and clinical outcome varies by adjuvant treatment, and tested the hypothesis that certain endome- trial cancer biomarkers will stratify progression-free survival (PFS) and overall survival (OS) in patients treated with (but not without) chemotherapy. Method: Genomic and clinical data from EEC patients treated without (no adjuvant treatment or radiation) versus with chemo- therapy (chemotherapy or chemotherapy and radiation) were downloaded from The Cancer Genome Atlas (TCGA). The rela- tionships between endometrial cancer biomarkers and clinical outcomes were investigated using log-rank testing and Cox hazards modeling. Results: There were 379 evaluable patients, including 92 (24%) who received chemotherapy. A panel of 15 endometrial cancer bio- markers were evaluated based either on mutation status and/or RNA sequencing. Transcript expression, not mutation, of CTNNB1 was the only biomarker that selectively stratified clinical outcome in EEC patients treated with chemotherapy. Patients with high versus low CTNNB1 had better PFS (Fig. 1A) and OS (Fig. 1B) following chemotherapy. The predictive value of CTNNB1 in chemotherapy patients was observed in univariate modeling (HR = 0.34, P = 0.001 for PFS; HR = 0.24, P = 0.025 for OS) and with multivariate adjustment for age of diagnosis, race, stage, grade, and myometrial invasion (HR = 0.27, P = 0.0009 for PFS; HR = 0.25, P = 0.039 for OS). As has been previously reported, high versus low CTNNB1 was associated with worse OS (Fig. 1D) in the lower risk EEC patients treated with no adjuvant treatment or radiation but not with chemotherapy, but this relationship was not associated with a corresponding increased risk of disease progression (Fig. 1C). Conclusion: EEC patients with high levels of CTNNB1 had better outcome following adjuvant chemotherapy with consistent improve- ments in PFS and OS. This relationship was not observed in patients who did not receive chemotherapy. Low transcript expression of CTNNB1 may be used to aid physicians in tailoring treatment choices to patients with EEC. doi:10.1016/j.ygyno.2017.03.191 164 - Poster Session Investigating variations in the SWI/SNF complex in serous ovarian cancer M.R. Davis a , W. Choi-Pui b , B. Howitt b , K. Hasselblatt c , R.S. Berkowitz c , S.W. Ng c . a Brigham and Women's Hospital/Harvard University, Boston, MA, USA, b Brigham and Women's Hospital/Harvard Medical School, Boston, MA, USA, c Brigham and Women's Hospital, Boston, MA, USA Objective: The SWI/SNF complex is a multi-subunit chromatin remodeling complex with tumor suppressor properties. Mutations in 4 subunits have been reported in ovarian clear cell carcinoma (OCCC). The aim of this study is to investigate the composition of the SWI/SNF complex and prognostic significance in epithelial ovarian cancer (EOC). Method: The Cancer Genome Atlas (TCGA) was queried to identify SWI/SNF mutations within the serous ovarian cancer dataset. Immunohistochemical (IHC) staining for SWI/SNF marker expression was performed on archived samples, and marker expression in EOC cell lines was determined by Western blot analysis. Complex composition was determined using mass spectrometric and Western blot analysis of co-immunoprecipitation (Co-IP) products. Survival analysis for SWI/SNF mutations was derived from data generated by the TCGA Research Network. Results: Among 316 cases of high-grade serous ovarian cancer (HGSC) in the TCGA dataset, 39% had mutations in the SWI/SNF complex. Thirty-four (11%) had amplification of SMARCA4, 52 (16%) had mutations in ACTL6A, while 2% had ARID1A mutations. Mutations in SMARCA4 were associated with a decreased overall survival (33.4 months vs 44.5 months, P = 0.04). On IHC, there was a strong correlation between expression of ARID2 and ARI1A and ARID1B (r = 0.78, P = 0.021 and r = 0.70, P = 0.035) and between Fig. 1. Progression-free survival (PFS) or overall survival (OS) plots were generated for endometrioid endometrial cancer patients women with low (blue line) vs high (red line ) transcript expression of CTNNB1 and treated with chemotherapy (A or B, respectively) vs without chemotherapy (C or D, respectively). Survival distributions were compared using log-rank test and univariate Cox modeling was performed estimating hazard ratio (HR) and 95% confidence interval (CI) for patients with high vs low transcript expression of CTNNB1. Abstracts / Gynecologic Oncology 145 (2017) 2–220 81