Abstract Background Adenocarcinoma of the gastro-esophageal junction (GEJ) is increasingly common in Western societies. GEJ tumours are typically categorized using the Siewert classification, which divides the tumours in to 3 types based on location relative to the GEJ. Studies have shown that there is improved survival with a proximal surgical resection such as esophagectomy for type 1 tumours while type 3 tumours have improved survival when managed as other gastric cancers with a total or extended total gastrectomy. This requires accurate clinical classification. This study aims to evaluate the accuracy of commonly used staging modalities in the clinical classification of GEJ tumours. Methods A review of 807 consecutive patients undergoing curative resection for GEJ cancer was performed. Patient data was obtained from a prospectively maintained database between February 2010 and April 2023. Siewert type was established clinically as part of the staging process. The standard staging modalities used for every patient included computerized tomography (CT) and endoscopy, with most patients also undergoing positron emission tomography (PET). Endoscopic ultra-sonography (EUS) and staging laparoscopy was also used selectively. Accuracy was determined by comparing clinical classification to pathological classification following surgery. Results 807 patients were included [319 type 1, 401 type 2, 87 type 3]. Clinical classification was the same as pathological classification in 59.5% of cases [69.6% for type 1, 48.6% for type 2, 71.3% for type 3 (p<0.001). The use of different staging modalities for each sub-type is presented in table 1. For patients that underwent EUS, Siewert type was accurate in 59.4% of cases overall. The failure rate for EUS was 2.9%. For patients undergoing staging laparoscopy Siewert type was accurate in 69.3% and for patients that underwent PET, EUS and staging lap the accuracy was 74.4%. Conclusion Accurately determining Siewert type clinically is challenging. This is particularly relevant for type 2 tumours which had the least accurate comparison of clinical to pathological type. The use of additional modalities including EUS, and staging laparoscopy improves accuracy but may not be feasible in all patients.