# Prevalence and endoscopic-histological correlation of premalignant gastric lesions at a university hospital in Uruguay

## TL;DR

This study examines the prevalence and endoscopic accuracy of detecting pre-cancerous stomach conditions in Uruguay, finding low correlation between endoscopy and biopsy results.

## Contribution

The study provides new prevalence data and evaluates endoscopic accuracy for pre-neoplastic gastric lesions in Uruguay.

## Key findings

- The prevalence of histological CAG and IM was 38.2% and 31.4%, respectively.
- Endoscopic-histological correlation for CAG and IM was low, with sensitivity of 46% and 22%, respectively.
- Interobserver agreement was good for certain endoscopic features like whitish-greyish plaques.

## Abstract

Although chronic atrophic gastritis (CAG), intestinal metaplasia (IM), and dysplasia constitute gastric pre-neoplastic conditions of gastric cancer (GC), data on endoscopic correlation and the prevalence in many South American countries are scarce. The aims of this study were to establish prevalence and perform endoscopic-histological correlation of gastric pre-neoplastic conditions using high-definition white light endoscopy (WLE) and to determine interobserver agreement for endoscopic findings for CAG and IM.

A prospective, observational, descriptive, cross-sectional study was carried out at a Uruguayan hospital during a 6-month period. Risk was stratified according to Operative Link for Gastritis Assessment and Operative Link for Gastric Intestinal Metaplasia stage for CAG and IM, respectively. An independent and blinded second observer was included to determine interobserver endoscopic and histologic agreement.

A total of 102 patients (mean age 57 years ± 1.6 years, 68.6% woman) were included.
Prevalence of histological CAG and IM were 38.2% and IM 31.4%, respectively.
Endoscopic-histological correlation for CAG had kappa index 0.063, sensitivity 46%, and
specificity 60%. For endoscopic IM, the kappa index was 0.216, sensitivity 22%, and
specificity 96%. Interobserver variability was good for gastric fold flattening and very
good in the presence of whitish-greyish plaques for CAG and IM, respectively.

The endoscopic-histological correlation of both CAG and IM was low, raising the need for biopsy for diagnosis in all cases, regardless of HD-WLE findings. Although prevalence of gastric pre-neoplastic conditions in this group of Uruguayan patients was comparable to those described in countries with a high incidence of GC, a low proportion of high-risk stages (III and IV) was identified.