Expression of SOX9 and CDX2 in nongoblet columnar-lined esophagus predicts the detection of Barrett's esophagus during follow-up.

The diagnosis of Barrett's esophagus in the United States requires both endoscopically evident columnar-lined esophagus and the presence of goblet cells by histology. Currently, there is no consensus on how patients with nongoblet columnar-lined esophagus should be followed. In this study, we investigated whether biomarkers can be used to predict the detection of goblet cells in follow-up biopsies. Patients with nongoblet columnar-lined esophagus were identified. In 13 of these cases, goblet cells were detected in subsequent follow-up endoscopic biopsies (Barrett's group). Additionally, 26 cases that remained negative for goblet cells in follow-up biopsies served as controls. Immunohistochemistry for CDX2, SOX9, BMP4, SHH, and MUC2 was performed on the initial biopsies and graded independently by at least two pathologists in a masked manner. CDX2 was positive in the nongoblet columnar epithelium of 7/13 cases in the Barrett's group and in 4/26 controls (sensitivity 54%, specificity of 85%, odds ratio (OR) 6.4). Strong and diffuse immunoreactivity for SOX9 was detected in 10/13 cases in the Barrett's group and in 1/26 controls (sensitivity 77%, specificity 96%, OR 83.3). Combining CDX2 and SOX9 as a panel increased sensitivity to 85%, although the specificity decreased to 85% (OR 30.3). SHH, BMP4, and MUC2 expression showed no significant difference between the Barrett's and control groups. In patients with nongoblet columnar-lined esophagus, SOX9 and CDX2 may be useful in identifying a subset of patients who have a higher risk of being diagnosed for Barrett's esophagus (developing goblet cells) and need closer follow-up.