Introduction Despite advances in the staging and treatment, the prognosis of upper gastrointestinal tract (UGIT) cancer in the UK remains poor, often presenting insidiously at a late stage. However, in contrast to our understanding of missed colorectal cancer rates following colonoscopy, relatively few studies have been published addressing the frequency of missed UGIT malignancies. Depending on the population studied this ranges from 6.7 to 25.8%. The aim of this study was to identify how frequently oesophagogastroduodenoscopy (OGD) may have failed to detect cancer at Prince Charles Hospital, a District General Hospital in South Wales, with a stable population of 150000, in the 36 months preceding a confirmed histological diagnosis.

Methods All patients between 1^st January 2010 and 31^st December 2012 who underwent an OGD and were subsequently diagnosed with cancer were identified using endoscopic records and the cancer service database. Patients who had undergone a prior endoscopy within 3 years of diagnosis were then identified and their records reviewed to analyse the previous endoscopic and histological findings.

Results 5454 endoscopies were performed during this time period, and a total of 134 patients (2.4%) with UGI cancer were identified. 77 (57%) were oesophageal, 49 (37%) gastric and 8 (6%) duodenal. The mean age was 69 years (range 24–91), with a higher proportion of males to females (3:1). Of these, 9 patients (6.7%) had undergone at least one previous endoscopy in the 36 months leading up to a confirmed UGIT cancer diagnosis, with 44% of these being within the preceding 12 months. The mean interval was 15months. The majority (55%) of patients had only one prior endoscopy (range 1– 5). 8 patients (89%) were found to have pathology at a preceding endoscopy at the site of a subsequently detected cancer. 6 patients were felt to have insufficient biopsy sampling (<4 or none) and 3 had inadequate surveillance or follow up of identified pathology (of which two had both inadequate sampling and surveillance).

Conclusion These findings, whilst similar to those previously reported in the literature have highlighted the importance of careful and thorough examination of the UGI tract, in particular with regard to adequate tissue sampling and surveillance. Consideration should be given to dedicated lists for surveillance of Barrett’s and the use of additional techniques such as narrow band imaging and chromoendoscopy in order to enhance diagnostic accuracy.

Disclosure of Interest None Declared.