The incidence of oesophageal adenocarcinoma has significantly increased in Europe over the last 30 years. The progression from normal mucosa to adenocarcinoma has been associated with genetic and morphological traits regrouped under the term "intraepithelial neoplasia" (IEN) according to the Vienna classification. The early detection of such lesions represents the first step in the identification of high-risk patients. The morphological criteria of IEN are the gold standard to identify such patients. Firstly described by Riddell et al in 1983, IEN is based on morphological criteria including both cytological and architectural alterations and is classified into different stages of severity. However, large studies have clearly demonstrated the lack of reproducibility, with large inter-individual discrepancies for both discrete and severe lesions. Discrepancies between high grade IEN and adenocarcinoma can be minimized by using the Vienna classification, which groups both of these lesions under the "stage IV". Discrepancies between low-grade IEN and uncertain lesions remain too important. Erroneous and overstated diagnosis of low grade IEN induces an unnecessary follow-up of patients with obvious psychological and economic consequences. Recent studies have demonstrated that the reading of the slides by 2 to 3 gastrointestinal (GI) pathologists significantly decreases interpretation mistakes. Because of these interpretation problems, scientists have looked for non-morphological criteria to confirm the pre-cancerous state. The current PubMed literature proposes many putative biomarkers. However, none of these has been correctly validated in large prospective case-control studies, which hampers their use in clinical routine. DNA quantification by flux cytometry and morphometry represent alternative methods of documenting IEN but these techniques are complex and expensive. The use of the proliferation marker Ki67 needs deep sampling with correct orientation and standardized cell counting. P504 S has been studied in Barrett's disease and might be a novel tool. The only promising tool thus far is the overexpression of p53 as shown in prospective studies demonstrating a nice correlation with clinical evolution and is easy to use in clinical routine.