This is a follow-up to my post from last month regarding columnar metaplasia of the esophagus. Dr. David Agbamu provided a reference to an article by renowned GI pathologists, Drs. Riddell and Odze, "Definition of Barrett's esophagus: time for a rethink--is intestinal metaplasia dead?" published online on July 21, 2009 in the Am J Gastroenterol.
I finally had a chance to read and digest this paper and heartily recommend practicing pathologist colleagues to do the same as I think it can inform (and reform?) our practices. GEJ biopsies and esophagus biopsies "r/o Barrett's" are common specimens seen in both "community" and "academic" practice settings.
Drs. Riddell and Odze make a compelling argument challenging the American College of Gastroenterology 2008 definition of BE, i.e., columnar-type mucosa at endoscopy (of any length) PLUS the identification of "intestinal metaplasia" (=goblet cells) on biopsy of the tubular esophagus. First, they contrast this definition with British and Japanese definitions of BE which do require any histologic confirmation. In brief, these definitions equate endoscopically recognized columnar-lined esophagus (CLE) with BE. Second, they outline evidence in support of dropping the requirement for goblet cells in the diagnosis of BE. The absence of goblet cells on biopsy of CLE is not a matter of sampling error and does not account for the dynamics of goblet cell expression (waxing/waning, conversion to nongoblet cell phenotype). Moreover, protein expression and molecular alterations are the same in CLE with and without goblet cells. Finally, there appears to be an equal risk of neoplastic progression in CLE regardless of goblet cells.
The clinical implications for pathologists is that it shifts the purpose of the index endoscopic examination and biopsy of CLE from the diagnosis of BE (Is there goblet cells or not?) to an initial surveillance for dysplasia (which is even more challenging!).
Finally, the authors argue for the fallacy of "ultrashort" BE: most agree that the endoscopic diagnosis of BE cannot be made with certainty if the CLE segment is < 1 cm. Of those biopsied, only a minority of these show histologic features that enable them to be differentiated as being of esophageal or gastric origin. But even if the origin of the biopsy can be determined, the risk of malignant progression of a segment of columnar metaplasia < 1 cm is unknown.
Again, this a real thought-provoking and well-documented article.