490. CORRELATION BETWEEN HISTOLOGICAL AND MOLECULAR ALTERATIONS WITH POOR SURVIVAL IN ESOPHAGEAL ADENOCARCINOMA

Abstract

Abstract The incidence of esophageal adenocarcinoma (EAC) has constantly risen in western countries over the past decades, often diagnosed at advanced stage and with a five-year survival around 20%1. In a previous study on EAC cases submitted to surgery (without neoadjuvant treatment) an algorithm (EACGSE classification)2 based on morphologic distinctions provided a significant prognostic impact. We aimed to evaluate the molecular basis underlying these differences, to improve patient management. The EAC cohort classified according to EACGSE2 was included in the study. Genomic DNA was available from formalin fixed paraffin embedded surgical specimens for 207 cases. The cases were sequenced for 26 cancer-related genes (panel #226722257; IDT) with high coverage on NextSeq500 (Illumina). Data analysis was performed using a dedicated pipeline3. 245 cases were analyzed for SMAD4 immunostaining (IHC). Loss of SMAD4 immunostaining was reported as % of negative tumor cells (at least 35% of neoplastic cells). TP53 was the most frequently altered gene (134/207 cases with at least one mutation). The presence of TP53 missense variants was associated to a poor cancer-specific survival (CSS) in the high-risk cases (glandular poorly differentiated, mucinous invasive, diffuse anaplastic, mixed) according to EACGSE2 (P=0.0005). A significant correlation was observed for TP53 truncative variants and loss of SMAD4 staining (P=0.008). Actually, SMAD4 loss was observed in 85/245 (35%) of all EAC cases analyzed via immunostaining. SMAD4 loss correlated with poor CCS (P=0.007) and disease-free survival (P=0.002) in EACGSE high-risk cases. TP53 missense mutations correlated to poor outcomes (CSS) in high-risk cases, TP53 truncative mutations were associated to SMAD4 loss. SMAD4 loss itself resulted a frequent event in EAC and correlated with lower CCS and disease-free survival in high-risk cases. Therefore, we were able to correlate EAC histological classification, clinical outcomes and molecular phenotypes. Validation in independent samples is warranted to corroborate these findings.