Early Barrett esophagus-related neoplasia in segments 1 cm or longer is always associated with intestinal metaplasia

• Published in: Modern pathology Vol. 30; no. 8; pp. 1170 - 1176
• Main Authors: Allanson, Benjamin Michael, Bonavita, Jessica, Mirzai, Bob, Khor, Tze Sheng, Raftopoulos, Spiro C, de Boer, Willem Bastiaan, Brown, Ian S, Kumarasinghe, Marian Priyanthi
• Format: Journal Article
• Language: English
• Published: New York Elsevier Inc 01-08-2017; Nature Publishing Group US; Elsevier Limited
• Subjects: 692/420/755; 692/499; Adenocarcinoma; Adenocarcinoma - pathology; Adult; Aged; Aged, 80 and over; Barrett Esophagus - pathology; Biopsy; Carcinoma; Dysplasia; Endoscopy; Esophageal Neoplasms - pathology; Esophagus; Female; Goblet cells; Humans; Intestine; Intestines - pathology; Laboratory Medicine; Male; Medicine; Medicine & Public Health; Metaplasia; Metaplasia - pathology; Middle Aged; Mucosa; Mucous membrane; original-article; Pathology; Sampling; Sampling error
• ISSN: 0893-3952; 1530-0285
• DOI: 10.1038/modpathol.2017.36

The assumption that intestinal metaplasia is a prerequisite for intraepithelial neoplasia/dysplasia and adenocarcinoma in the distal esophagus has been challenged by observations of adenocarcinoma without associated intestinal metaplasia. This study describes our experience of intestinal metaplasia in association with early Barrett neoplasia in distal esophagus and gastroesophageal junction. We reviewed the first endoscopic mucosal resection of 139 patients with biopsy-proven neoplasia. In index endoscopic mucosal resection, 110/139 (79%) cases showed intestinal metaplasia. Seven had intestinal metaplasia on prior biopsy specimens and three had intestinal metaplasia in subsequent specimens, totaling 120/139 (86%) patients showing intestinal metaplasia at some point supporting the theory of sampling error for absence of intestinal metaplasia in some cases. Those without intestinal metaplasia (13%) were enriched for higher stage disease (T1a Stolte m2 or above) supporting the assertion of obliteration of intestinal metaplasia by the advancing carcinoma. All cases of intraepithelial neoplasia and T1a Stolte m1 carcinomas had intestinal metaplasia (42/42). The average density of columnar-lined mucosa showing goblet cells was significantly less in shorter segments compared to those ≥3 cm (0.31 vs 0.51, P=0.0304). Cases where segments measured less than 1 cm were seen in a higher proportion of females and also tended to lack intestinal metaplasia. We conclude that early Barrett neoplasia is always associated with intestinal metaplasia; absence of intestinal metaplasia is attributable to sampling error or obliteration of residual intestinal metaplasia by neoplasia and those with segments less than 1 cm show atypical features for Barrett-related disease (absent intestinal metaplasia and female gender), supporting that gastroesophageal junction adenocarcinomas are heterogeneous.