Patients with gastric adenocarcinoma have a poor prognosis. One factor contributing is advanced disease at diagnosis. Chronic atrophic gastritis (CAG) progresses in a stepwise fashion from chronic gastritis, gastric atrophy (GA), gastric intestinal metaplasia (GIM) to gastric dysplasia and ultimately gastric adenocarcinoma. 

The British Society of Gastroenterology (BSG) and European Society of Gastrointestinal Endoscopy (ESGE) both have guidelines aiming to improve the diagnosis and management of patients at high risk of gastric adenocarcinoma^(1,2). 

They advise gastroscopy with high quality mucosal visualisation and biopsies for histopathology to detect GA/GIM and allow appropriate risk stratification. GIM is endoscopically visualised by whitish, patchy, and irregular mucosal areas with a potentially velvety or slightly nodular surface. When identified, Sydney protocol biopsies should be completed (biopsies from the antrum, incisura, the lesser and greater curvatures). 

High risk individuals with extensive or advanced histological stages of CAG/GIM should undergo 3-yearly endoscopic surveillance as per guidelines^(1,2). This also includes people with a family history of gastric cancer or persistent Helicobacter pylori (H. pylori) infection. 

The aims of this audit were to analyse:

1. How well GIM is visualised endoscopically, and if Sydney protocol biopsies are followed? 

2. When GIM is confirmed histologically, is surveillance organised and is this at the correct frequency?

We carried out a retrospective audit of endoscopy data at Barnet General, Chase Farm and Royal Free Hospitals, London, U.K., from June to October 2025. Patients were identified from a histology database of specimens with GIM and then data scrutinized using endoscopy reporting software and electronic patient records.

Between June and October 2025, 4394 gastroscopies were performed. 145 patients were confirmed to have GIM. 47/145 records were excluded, coming under either neuroendocrine tumour or Barrett’s oesophagus surveillance or records with incomplete data. This left 98 endoscopies that were fully analysed. 58.2% were female with a mean age of 67. 11 of the patients were H. pylori positive.

GIM was visualised endoscopically in 19/98 (19.4%) cases, being confirmed histologically in 100% of them. In 79/98 (80.6%) cases GIM was histologically confirmed, however, not macroscopically visualised during gastroscopy. Of the 19 records where GIM was visualised, only 13/19 (68.4%) had Sydney protocol biopsies taken. Sydney protocol biopsies were also performed in 15.3% of cases where GIM/CAG were not visualised. 

Extensive/corpus GIM accounted for 42/98 (42.9%), antrum/incisura 32/98 (32.7%), not specified biopsies 24/98 (24.5%). For those with extensive/corpus GIM, surveillance was correctly organised in 19/42 (45.2%) of the cohort. However, this was not always 3-yearly and ranged between 6 months to 3 years. For those with antrum/incisura GIM, 25/32 (78.1%) correctly had no surveillance organised, however, for the remainder it was still being organised.

In 1/98 patient, gastric biopsy showed GIM with ulceration. Repeat gastroscopy with biopsies and histology confirmed a poorly differentiated gastric adenocarcinoma.

1. Our results indicate that GIM was endoscopically visualised in only 20% of cases. 

2. Even when identified correctly, Sydney protocol biopsies are not always completed. 

3. Large variation exists in whether surveillance is being organised for patients with extensive/corpus GIM and the frequency of this surveillance.

4. >20% of patients with limited GIM are having surveillance organised unnecessarily. 

We recommend better training in the endoscopic detection of GIM and raising awareness of the surveillance guidelines. By ensuring the correct biopsies are taken, patients may benefit from earlier detection and management of pre-cancerous lesions and avoid unnecessary procedures. This would also help improve efficiency and costs within our endoscopy services.