244. A NOVEL SERUM GLYCOPROTEIN BIOMARKER PANEL FOR SCREENING OF ESOPHAGEAL ADENOCARCINOMA AND SURVEILLANCE OF BARRETT’S ESOPHAGUS

Abstract

Abstract Background Esophageal adenocarcinoma (EAC) has become the predominant type of esophageal cancer in Western countries and Barrett’s esophagus (BE) is the only known precursor of EAC. Early diagnosis and intervention of EAC significantly reduces the disease mortality. Endoscopic surveillance in patients with BE is costly and remains controversial as an EAC screening test at the population level. This study aims to develop a simple blood test to allow close surveillance of EAC in patients with BE and screening EAC at a large population level. Methods A total of 298 serum samples were obtained from two independent cohorts recruited in Australia (N = 249) and the US (N = 49) from four groups: healthy controls, BE-low/no dysplasia, BE-high dysplasia and EAC. Thirty-three candidate glycoprotein peptides were measured by mass spectrometry after lectin pulldown. The Australian cohort was used as the development cohort. Multivariate logistic regression was used to develop diagnostic models to distinguish different disease status, followed by a forward stepwise approach. Diagnostic models were further validated in the US cohort. Model performance was assessed by the area under the receiver operating characteristics curve (AUC-ROC) and sensitivity, specificity of each model determined. Results Seven glycoproteins were found to be significantly associated with EAC. A series of diagnostic models were developed to distinguish different disease status, with AUCs ranging from 0.72 to 0.84 in the development cohort and from 0.79 to 0.94 in the validation cohort. After adjusting for age and sex, biomarker models showed good discrimination and can distinguish patients with EAC from control group (AUC = 0.75 in development cohort and AUC = 0.94 in validation cohort). Biomarker models can also distinguish patients who require endoscopic interventions (including patients with EAC and BE with high dysplasia) from participants in the control group (AUC = 0.72 in development cohort and AUC = 0.79 the validation cohort). Conclusion A panel of serum glycoprotein biomarkers were identified associated with EAC and BE with high dysplasia, with good discrimination of disease status. These biomarkers have potential utility as a simple blood test that could be used for screening EAC in a large population, and for surveillance disease progression in patients with BE.