Oesophageal cancer is a global health problem with high mortality due to the advanced nature of the disease at presentation; therefore, detection at an early stage significantly improves outcome. Oesophageal squamous-cell cancer is preceded by dysplasia and oesophageal adenocarcinoma is preceded by Barrett's oesophagus, which progresses to cancer via intermediate dysplastic stages. Screening to detect these preneoplastic lesions has the potential to substantially reduce mortality and morbidity. However, the risks and benefits of such programmes to individuals and to society need to be carefully weighed. Endoscopic screening is invasive, costly and error prone owing to sampling bias and the subjective diagnosis of dysplasia. Non-endoscopic cell-sampling methods are less invasive and more cost effective than endoscopy, but the sensitivity and specificity of cytological assessment of atypia has been disappointing. The use of biomarkers to analyse samples collected using pan-oesophageal cell-collection devices may improve diagnostic accuracy; however, further work is required to confirm this. The psychological and economic implications of screening as well as the feasibility of implementing such programmes must also be considered.