Background: It is unclear if differential chemotherapy effects exist on overall survival (OS), response rate (RR) and toxicity depending on primary tumour origin [oesophageal versus oesophago-gastric junction (OGJ) versus gastric adenocarcinoma].
Patients and methods: A total of 2110 patients were enrolled in four randomised controlled trials (RCTs) assessing fluoropyrimidine +/- platinum-based chemotherapy. This analysis used individual patient data and restricted to patients with adenocarcinoma who received one or more dose of chemotherapy. Gastric origin was the control in comparisons of tumour origin.
Results: Of the 2110 patients randomised, 1775 (84%) patients had adenocarcinoma with oesophageal (n = 485), OGJ (n = 457) and gastric (n = 833) origins. The median OS was 9.5 months in oesophageal, 9.3 months in OGJ and 8.7 months in gastric cancer (P = 0.68). RR was 44.1% in oesophageal, 41.1% in OGJ and 35.6% in gastric cancers (P = 0.11 and 0.27, respectively, compared with gastric cancer on multivariate analysis). Toxicity composite end point occurred in 46%, 47% and 45% in oesophageal, OGJ and gastric cancers, respectively (P = 0.85 and 0.62 compared with gastric).
Conclusions: In our large multicentre RCT dataset, no significant differences were demonstrated on multivariate analyses in OS, RR and toxic effects among patients with advanced oesophageal, OGJ and gastric adenocarcinoma. Future RCTs should not exclude oesophageal adenocarcinoma.