Endoscopic features of buried Barrett's mucosa.

BACKGROUND AND AIMS: Buried Barrett's mucosa is defined as intestinal metaplasia that is "buried" under the normal-appearing squamous epithelium. This can occur in Barrett's esophagus with or without previous endoscopic therapy. Dysplasia and neoplasia within buried Barrett's mucosa have also been reported. However, endoscopic features of buried Barrett's mucosa have not been described. At our tertiary referral center for Barrett's esophagus, several endoscopic features have been observed in patients who were found to have buried Barrett's mucosa on histology. These features are squamous epithelium which is (1) darker pink on white-light and darker brown on narrow-band imaging and/or (2) has a slightly raised or nodular appearance. It was also observed that either of these 2 features is frequently seen adjacent to a Barrett's mucosa island. This study aimed to (1) evaluate the diagnostic accuracy of these endoscopic features, and (2) evaluate the frequency of endoscopically identifiable buried Barrett's mucosa in patients with dysplastic Barrett's esophagus, before and after endoscopic eradication therapy.

METHODS: This was a retrospective analysis of a prospectively observed cohort of all cases of dysplastic Barrett's esophagus referred to St Vincent's Hospital, Melbourne. Endoscopy documentation software and histopathology reports of esophageal biopsy and EMR specimens between March 2013 and March 2019 were searched for terms "buried" or "subsquamous" Barrett's mucosa. Endoscopic reports, images, and histopathology reports of suspected buried Barrett's mucosa were then reviewed to apply the endoscopic features and correlate with the histologic diagnosis.

RESULTS: In a cohort of 506 patients with dysplastic Barrett's esophagus, 33 (7%) patients (73% male, median age at referral 70.5 years) had buried Barrett's mucosa on histology. Twenty-seven (82%) patients had previous treatment for dysplastic Barrett's esophagus; radiofrequency in 2 (6%), EMR in 4 (12%), and both modalities in 21 (64%). Six (18%) had no previous treatment. Histologically confirmed buried Barrett's mucosa was suspected at endoscopy in 26 patients (79%). Endoscopic features were (1) darker pink or darker brown mucosa underneath squamous epithelium (24%), (2) raised areas underneath squamous mucosa (27%), and both features present concurrently (27%). These features were associated with adjacent islands of Barrett's esophagus in 48%. Forty-four cases of buried Barrett's mucosa were suspected endoscopically, and these were sampled by biopsy (50%) and EMR (50%). Buried Barrett's mucosa was confirmed in 26 cases, with a positive predictive value of endoscopic suspicion of 59%. Eighteen cases of endoscopically suspected buried Barrett's mucosa had no buried Barrett's mucosa on histology; inflammation or reflux was identified in 12 (67%) patients. Dysplasia was identified within buried Barrett's mucosa in 12 (36%) patients; 5 intramucosal adenocarcinoma, 1 high-grade dysplasia, and 6 low-grade dysplasia. Endoscopic features of buried Barrett's mucosa were observed in 11 of 12 cases harboring dysplasia or neoplasia, compared with 15 of 21 cases of buried Barrett's mucosa without dysplasia.

CONCLUSIONS: In this retrospective analysis of prospectively observed patients with dysplastic Barrett's esophagus, buried Barrett's mucosa was identified in 7%, including treatment-naive patients. The proposed endoscopic features of buried Barrett's mucosa were seen in 79% of patients with histology confirmed disease. These endoscopic features may predict the presence of buried Barrett's mucosa, which may contain dysplasia or neoplasia. An overlap between the endoscopic features of inflammation, reflux, and buried Barrett's mucosa was observed. Future prospective studies are required to develop and validate endoscopic criteria for identifying buried Barrett's mucosa.