Outcomes of patients with Barrett's oesophagus with low-grade dysplasia undergoing endoscopic surveillance in a tertiary centre: a retrospective cohort study

Abstract

Barrett's oesophagus predisposes individuals to oesophageal adenocarcinoma (OAC), with the risk of progression to malignancy increasing with the degree of dysplasia, categorized as either low-grade dysplasia (LGD) or high-grade dysplasia (HGD). The reported incidence of progression to OAC in LGD ranges from 0.02% to 11.43% per annum. In patients with LGD, Australian guidelines recommend 6-monthly endoscopic surveillance. We aimed to describe the surveillance practices within a tertiary centre, and to determine the predictive value of surveillance as well as other risk factors for progression. Endoscopy and pathology databases were searched over a 10-year period to collate all cases of Barrett's oesophagus with LGD. Medical records were reviewed to document patient factors and endoscopic and histologic details. Because follow-up times varied greatly, survival analysis techniques were employed. Fifty-nine patients were found to have LGD. Thirteen patients (22.0%) progressed to either HGD or OAC (10 (16.9%) and three (5.1%) respectively); the annual incidence rates of progression to HGD/OAC and OAC were 5.5% and 1.1% respectively. All patients who developed OAC had non-guideline-adherent surveillance. A Cox model found only two predictors of progression: (i) guideline-adherent surveillance, performed in 16 (27.1%), detected progression to HGD/OAC four times earlier than non-guideline-adherent surveillance (95% confidence interval (CI) = 1.3-12.3; P = 0.016). (ii) The detection of visible lesions at exit endoscopy independently predicted progression (hazard ratio = 6.5; 95% CI = 1.9-22.8; P = 0.003). Barrett's oesophagus with LGD poses a significant risk of progression to HGD/OAC. Guideline-recommended surveillance is effective, but is difficult to adhere to. Clinical predictors for those who are more likely to progress are yet to be defined.