4178 Background: Barrett esophagus is the precursor of most esophageal adenocarcinomas. Dysplasia or intraepithelial neoplasia (IEN) is the first step in the sequence which can be recognized with routine morphology in endoscopic biopsies. This diagnosis is however difficult in biopsies with high interobserver variability for low-grade IEN. To solve this diagnostic challenge, immunohistochemistry for p53, a tumor-supressor marker, has been proposed. The features which are useful for assessment of p53-immunoreactivity (p53+) and their clinical and pathogenetic implications are still unclear. Aim: To evaluate the relationship between p53+ and morphological changes and Ki-67-positivity (Ki-67+), a proliferation marker, in Barrett mucosa. Methods: Histology and immunohistochemistry (p53, clone DO-7;Ki-67, clone Mib-1) was performed on 20 operation specimens (short segment Barrett n=10, long n=10) (M/F 3, 66±10 yr) (IEN: low grade n=3, high grade n=3, adenocarcinoma: in situ n=2, T1 n=12; 100% of lesions p53+). p53+ in non-tumoral Barrett mucosa (focal i.e. one or more foci of positive glands vs. diffuse) and its localization (surface epithelium vs. basis of crypts) was analyzed in function of architectural and cytological features of IEN (nuclear enlargement, pseudostratification, maturation). Results: p53+ was detected in 90% of cases (focal: 45%; diffuse: 45%). Diffuse p53+ correlates with carcinoma. Five cases did contain 1 or 2 p53+ foci, 4 cases 4 up to 22 foci. No association was found with architectural abnormalities. A significant difference (p<0.0001)was observed in nuclear enlargement (18% vs. 53%), pseudostratification (11% vs. 79%) and absence of maturation of the epithelium (8% vs. 59%). Focal p53+ was associated with clear-cut morphological IEN in 17% of foci, with a significant lower sensitivity (20% vs. 80%), specificity (27% vs. 73%) and diagnostic accuracy (24% vs 76%) than diffuse p53+ (p<0.0001). Ki-67+ did not improve the diagnostic accuracy. p53+ surface epithelium equals IEN. Conclusion: p53+ was found in two patterns (focal vs. diffuse), whereby the latter correlates to morphological IEN and the former may be indicative of early changes which are not yet detectable by routine morphology. No significant financial relationships to disclose.