With interest we read the article by Hansen et al. (Gut 2007;56:918–25). The authors have to be congratulated on their contribution which profoundly adds to our understanding of the pathophysiology leading to oesophageal and gastric cancer. In addition, the paper points out the difficulties we have to correctly assign tumours of the oesophago-gastric junction (oesophageal vs gastric?). Cardiac cancers were subtyped for their associations with serum anti-Helicobacter pylori IgG antibody titer and biochemical markers of loss of gastric secretory function associated with atrophy (pepsinogen I/II ratio and serum gastrin concentration). However, the anatomic criterium to define cardiac carcinoma, that is, tumours centred within 2 cm distal to the oesophago-gastric junction, is inaccurate. Interpretation of the data should be conducted with the inclusion of clear, anatomical and histopathological criteria.

It is well accepted that cardiac cancer and adenocarcinoma of the oesophagus share epidemiological and pathogenetic features.^1,2 After birth the oesophagus is lined by squamous epithelium, whereas the stomach is lined by gastric oxyntic mucosa (with parietal and chief cells).^1,3 Due to gastro-oesophageal reflux, squamous epithelium is …