Abstract 5410: An miRNA signature selects patients at risk for Barrett's Esophagus progression to dysplasia and cancer

Background: Esophageal adenocarcinoma (EAC) is an aggressive malignancy with an increasing incidence in the US. Progression of Barrett's Esophagus (BE) to EAC occurs via a stepwise process, and consequently periodic esophageal biopsies are utilized in order to monitor patients with BE. While a small percent of BE patients will progress to dysplasia and cancer, the majority of them will continue to have long-standing BE without progression. In the absence of dysplasia, the behavior of BE cannot be predicted based on evaluation of histologic features alone. Studies have demonstrated the utility of microRNAs in differentiating between the specific evolutionary events in the progression of BE to dysplasia and cancer. To date, there has yet to be a comparison between cases of BE that have not progressed to dysplasia or carcinoma, or cases of BE-Non-progression (BEN), and cases of BE that have progressed to dysplasia and/or carcinoma, or cases of BE- Progression (BEP). Through miRNA profiling, we have determined an assay of candidate miRNAs that reliably differentiate cases of BEN from BEP. Methods: Fifty cases of BE were profiled with two different miRNA profiling techniques: the HTG EdgeSeq miRNA WT Assay and the nanoString Assay. There were 15 cases of BEN (follow up >7 years) as well as of 11 cases of BEP (progression to dysplasia and/or EAC within 3 years). These cases were used for miRNA discovery and miRNA prediction model training. Another independent patient data set of 24 BE cases (13 cases of BEN and 11 cases of BEP; namely Normalized Nano), was profiled with nanoString miRNA Assay. This set was used as an independent validating cohort. Results: Six significant miRNAs were identified and confirmed using two different statistical methods (Limma test and Wilcoxen rank sum test). Family-wise error rate for type I error (FWER) was set to less than 0.05 in order to minimize the probability of discovering false positives. The final miRNA model demonstrated a high prediction performance at the optimal cutoff with a specificity of 80% and 50%; as well as sensitivity 100% and 75% for the two sets in order to capture a high proportion of the BEPs. We then independently validated this miRNA signature with the Normalized Nano set, which demonstrated a sensitivity of 70% and a specificity of 67.5% in this independent validation. While the prediction performance was weaker, we consistently validated its prediction power and clinical utility on an independent patient cohort with a completely different miRNA profiling technique. Conclusion: The reliability of these candidate miRNAs support further investigation in a larger population of patients, and may have potentially prognostic utility in the evaluation of BE patients with the goal of early detection of BE progression. Citation Format: James Saller, Kun Jiang, Kevin Neill, Zachary Mayer, Jae Lee, Luis Pena, F. Scott Corbett, Jose Pimiento, Mokenge Malafa, Anthony Magliocco, Domenico Coppola. An miRNA signature selects patients at risk for Barrett's Esophagus progression to dysplasia and cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5410.