473 CORRELATION OF FLUORESCENCE IN SITU HYBRIDIZATION (FISH) FINDINGS AT ERCP WITH DIAGNOSIS OF BILIARY MALIGNANCY IN PATIENTS WITH AN INDETERMINATE BILIARY STRICTURE: A CASE- CONTROL STUDY Scott H. Coppel, Mark A. Gromski, Gail H. Vance, Jeffrey J. Easler, James L. Watkins, Evan L. Fogel, Benjamin L. Bick, Glen A. Lehman, Stuart Sherman Introduction: Indeterminate biliary strictures present a clinical diagnostic challenge, given the paucicellular nature of biliary-specific malignancies. Multiple diagnostic modalities are utilized to aid in the diagnosis of indeterminate biliary strictures. To date, the value of fluorescence in situ hybridization (FISH) from biliary brushings at the time of endoscopic retrograde cholangiopancreatography (ERCP) has not been clearly defined. The aim of this study is to evaluate the correlation of specific FISH findings from ERCP biliary brushings in patients who do and do not have proven biliary malignancies. Methods: This is a case- control study from a single high-volume ERCP center. Patients were identified using a prospectively collected procedural database from 3/12/10 to 4/24/18. Patients who underwent ERCP for an indeterminate biliary stricture which included FISH sampling who had a subsequent confirmed tissue diagnosis of adenocarcinoma from the biliary stricture (e.g., surgical specimen, fluoroscopic or site directed biopsies and/or confirmed adenocarcinoma on cytology brushings) were considered cases. Each patient with confirmed adenocarcinoma (cases) was matched with a patient with an indeterminate biliary stricture who had undergone ERCP with FISH sampling for which there was no confirmation of adenocarcinoma related to the biliary stricture within the immediate 12 month period post-index ERCP. Matching of cases and controls was performed for age (+/- 2 years) and primary sclerosing cholangitis (PSC) status. Results: During the study period, 105 patients were identified to meet criteria for the confirmed adenocarcinoma group (cases). These were matched with 105 control patients who had no evidence of related adenocarcinoma within 12 months of the ERCP with FISH sampling, matched for age and PSC status. 192 patients had no evidence of PSC and 18 had PSC. Abnormal FISH findings of polysomy of chromosome 3, 7 and 17 (individ- ually and in combination) and also deletion of 9p21.3 were significantly more common in the cases with adenocarcinoma than the controls (all patients – Table 1). This was consistent for all patients and for the non-PSC cohort. For the PSC cohort (n=9 each group), although there were numerical trends consistent with the above, none of the FISH findings met statistical significance. Discussion: FISH findings of polysomy of chromosome 3, 7 and 17 and also deletion of 9p21.3 from indeterminate biliary strictures obtained at ERCP were significantly more common in the patients with adenocarcinoma than in controls. This suggests that polysomy 3, 7, 17 and/or 9p21.3 deletion by FISH should be considered concerning, and if other concurrent sampling modalities do not identify adenocarcinoma, that follow-up imaging and/or ERCP with sampling should be performed at a short interval. Surgery may also be considered in select cases. S-91 AGA Abstracts 474 CORRELATES OF ULTRASOUND QUALITY FOR HCC SURVEILLANCE IN PATIENTS WTIH CIRRHOSIS Haley Schoenberger, David Fetzer, Takeshi Yokoo, Nicole E. Rich, Ioana Tihina, Mobolaji Odewole, Amit G. Singal Background: Abdominal ultrasound fails to detect over one-fourth of hepatocellular carci- noma (HCC) at an early stage in patients with cirrhosis. Identifying patients in whom ultrasound is inadequate to exclude the presence of HCC can inform interventions to improve surveillance effectiveness. We aimed to evaluate and identify correlates of inadequate ultrasound quality in patients with cirrhosis undergoing HCC surveillance. Methods: We performed a retrospective cohort study of patients with cirrhosis who underwent ultrasound examination at a large urban safety-net health system between July 2016 and July 2019. Patients were initially identified using ICD-10 codes, with presence of cirrhosis then con- firmed by manual chart review. Ultrasound exam adequacy was defined by the interpreting radiologists routine clinical practice using LI-RADS visualization score (A=no/minimal limita- tions, B=moderate limitations, C=severe limitations). For patients with >1 ultrasound exam, we characterized change in visualization scores over time. We performed multivariable logistic regression to determine characteristics associated with inadequate ultrasound visualiz- ation (LI-RADS score B or C). Results: We identified 411 cirrhosis patients with at least 1 ultrasound exam, with most being an outpatient exam (78.6%) for surveillance indication (73.5%). Most (78.8%) had Child A cirrhosis, with only 17.5% having ascites or hepatic encephalopathy. Overall, 355 (86.4%) were classified as score A, 43 (10.4%) score B, and 13 (3.2%) score C, with the most common limitation being heterogeneous echotexture. In multivariable analysis, inadequate ultrasound visualization was associated with Child B or C cirrhosis (OR 2.80, 95% CI 1.33-5.91) and presence of obesity (OR 2.80, 95%CI 1.04- 7.53 for overweight; OR 2.36, 95%CI 0.81-6.83 for obesity; OR 11.0, 95%CI 3.69-32.7 for morbid obesity). Inpatient status was associated with inadequate visualization on univariate analysis but was not significant in multivariable analysis. Ultrasounds were inadequate in 36.6% of patients with morbid obesity and 24.7% of patients with Child B/C cirrhosis. Among 318 patients with >1 ultrasound exam, 250 (78.6%) had the same visualization score (234 score A, 13 score B, 3 score C). However, 44 (15.8%) of 278 patients with score A had inadequate visualization on repeat exam (42 score B, 2 score C) and 18 (45%) of 40 patients with an initial inadequate ultrasound had score A when repeated. Changes in visualization score were consistent across subgroups including patients with obesity or Child Pugh B-C cirrhosis. Conclusions: Over 10% of ultrasound exams are of inadequate quality for HCC surveillance, particularly among patients with obesity or Child Pugh B-C cirrhosis. However, ultrasound quality can change between exams, including improvement noted in half of patients with an initial inadequate ultrasound. 475 CROTONYLATION OF SEPT2 PROMOTES METASTASIS AND INVASIONS IN HEPATOCELLULAR CARCINOMA THROUGH STABILIZING P85 ALPHA Ning Zhang, Zhang Xinyue, Xiaoxing Li, Lixia Xu, Minhu Chen Background & aims: Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer death and the tumor metastasis and invasion are restriction in the therapy. The complexity of biology not only depends on genome but also the proteome such as posttranslational modification (PTM), which affects protein properties and relates to the therapeutic and prognostic applications. Lysine crotonylation (Kcr) is a new PTM found in 2011. But the function of Kcr in proteins is still unclear. In this study, we explored the role of non-histone Kcr in metastasis of HCC. Methods: Kcr level of proteins in the cell lines MHCC97H (high invasive) and MHCC97L (low invasive) were detected by stable isotope labeling with amino acids in cell culture (SILAC). The differentially crotonylated proteins of two cell lines was selected by bioinformatic analysis. Kcr of proteins were confirmed by immunoprecipitation and followed by western bolt. Protein-protein interaction was detected by GST-pull down assay and co-immunoprecipitation assay. The function in metastasis of Kcr were examined by Transwell migration and invasion assay. Results: By using SILAC, we found 334 upregu- lated and 205 downregulated proteins (Fold change > 1.5, P < 0.05) in MHCC97H, compared with MHCC97L. And SILAC showed that there are 199 up-crotonylated and 46 down- crotonylated sites (Fold change > 1.5, P < 0.05) in MHCC97H. After normalized by protein levels, there are 10 down-crotonylated sites and 20 up-crotonylated sites (Fold change > 1.5, P < 0.05). We detected the Kcr level of 3 proteins at the top list of high-Kcr but no apparently expressing changes. The results showed that candidates are cortonylated in cells and hyper-cortonylated in MHCC97H. We chose SEPT2, had the most difference in Kcr of 3 proteins between MHCC97H and MHCC97L, to perform further functional assays. After screening from known crotonylases and decrotonylases, we found P300 crotonylated SEPT2 and SIRT2 decrotonylated SEPT2. SILAC showed SEPT2 was hyper-crotonylated at lysine 74 (K74) and 318 (K318) in high invasive cells. According to structure analysis and species conservation of SEPT2, we conjectured that K74 was the main functional crotonylation site. A single mutation at K74 (K74R) inhibited invasion and metastasis of HCC cells in vitro. Protein interaction assays showed that K74R can suppress the interaction with P85 α,a subunit of PI3K complex, and SEPT2. Western blot showed that K74R facilitated P85 α degradation, suppressed PI3K pathway, and inhibited epithelial–mesenchymal transition pathway. Conclusions: In this study, we firstly explored the role of Kcr on non-histone protein and therefore regulating metastasis and invasion of cancer. SEPT2, hyper-crotonylated in high invasive cells, regulated by SIRT2 and P300, can promote metastasis and invasion in HCC through stabilizing P85 α and facilitating PI3K pathway. AGA Abstracts