Sa1867 Evidence of Two Aetiologies of Gastroesophageal Junctional Cancers Based on Gastric Parietal Cell Density

Introduction Serum pepsinogen I:II ratio, a surrogate marker of atrophic gastritis, suggests that some adenocarcinomas at the gastroesophageal junction (GOJ) develop on a background of atrophic gastritis, similar to non-cardia gastric cancer, while others arise on a backgrounds of healthy, non-atrophic gastric mucosa similar to oesophageal adenocarcinoma. In this current study, we have directly the background gastric body mucosa in patients with junctional adenocarcinomas compared to oesophageal adenocarcinomas and non-cardia gastric cancers. Methods 127 gastrectomy and oesophagectomy specimens for adenocarcinoma were identified for which clear topographic description allowed assignment to oesophageal, junctional (including cardia) and gastric non-cardia locations. In these gastric body mucosa specimens, well clear of the tumour margin, parietal cells were immunostained using anti- H + /K + ATPase. Parietal cell density was counted in 3 to 5 well-oriented fields (1 mm 2 each) and expressed as mean parietal cell number per 1 mm 2 area. Total mucosal thickness, glandular thickness, intestinal metaplasia, inflammation indicated by polymorphonuclear (PMN) and mononuclear (MN) cells and reactive atypia (RA) were also scored. Non-parametric statistics were used to compare distributions. Results Ten (8%) cases lacked well-orientated blocks of body mucosa. The remaining 117 patients included 34 oesophageal, 52 GOJ and 31 non-cardia gastric adenocarcinomas. Median (IQR) parietal cell densities were 836 (173), 602 (389) and 411 (334) per mm 2 in gastric mucosa of oesophageal, GOJ and gastric cancers, respectively (all differences P 2 , 85% of oesophageal adenocarcinomas had a higher and 84% of non-cardia gastric cancers had a lower values. With the same cut-off, 50% of GOJ adenocarcinomas were gastric and remaining was oesophageal in origin. Glandular mucosa was thicker in patients with GOJ cancer compared to gastric (0.735 vs. 0.600, p = 0.005) and thinner than oesophageal cancer (0.735 vs. 0.900, p Conclusion This study provides direct evidence for marked differences in the gastric mucosal phenotype in the patients with oesophageal versus gastric non-cardia cancer, with the former being healthy and uninflamed, but the latter atrophic and inflamed. The background gastric mucosa of GOJ cancer supported them being two distinct aetiologies, one group resembling oesophageal adenocarcinoma and other gastric non-cardia cancer. Disclosure of Interest None Declared.