Polypoid dysplasia in Barrett's esophagus: a clinicopathologic, immunohistochemical, and molecular study of five cases.

Dysplasia in Barrett's esophagus (BE) is a precursor to adenocarcinoma and most commonly occurs as a flat, grossly undetectable lesion. Rarely, dysplasia in BE may grow as a polypoid lesion. Most BE-associated polypoid dysplastic lesions have been referred to as "adenomas" because of their histological similarity to a colonic adenoma. BE-associated polypoid dysplastic lesions have been less well characterized than the flat type. Therefore, our aim was to characterize the clinicopathologic and molecular features of five cases of BE-associated polypoid dysplasia and to review the literature on this entity. The cases were evaluated clinically, histologically, immunostained for MIB-1 and p53, and genotyped for loss of heterozygosity (LOH) at the adenomatous polyposis coli (APC) locus. Mucosal biopsy specimens of five BE patients without dysplasia, and five BE cases with high-grade flat dysplasia, were used as controls. The study patients were all male (average age, 71 years) who presented with symptoms of gastroesophageal reflux disease. Endoscopically, all five cases had a well-defined sessile or pedunculated polypoid lesion ranging from 0.4 to 1.5 cm in size in the mid (n = 1) or distal (n = 4) esophagus and were associated with specialized-type BE (four long segment, one short segment). Histologically, the polyps consisted of intestinalized epithelium with low- and high-grade dysplasia. All five cases contained adenocarcinoma (four within the polyp, one in adjacent BE). All polyps showed increased cell proliferation in the form of surface MiB-1 staining and showed positive p53 staining. Three of three (100%) informative cases showed LOH at the APC locus in the dysplastic epithelium and in areas of adenocarcinoma. All five flat dysplasia controls also showed surface MIB-1 staining and p53 positivity, and three of three informative controls showed LOH for APC. None of the nondysplastic BE controls showed any of these findings. Three patients were treated with esophagectomy and two with polypectomy. All were alive, without metastasis, from 2 months to 6 years later. A literature review of esophageal "adenomas" uncovered 12 cases. Four of these had no clinical or pathological information, two were, in fact, gastric heterotopic lesions, one was composed entirely of intestinal-type epithelium, and five were polypoid dysplastic lesions similar to the cases described here (three male, two female; mean age, 59 years). Four of these five cases were associated with adenocarcinoma in the polyp (two intramucosal, two submucosal). In summary, BE-associated polypoid dysplasia share similar clinical, pathological, and molecular features as flat dysplasia and are often associated with adenocarcinoma. Thus, we agree with other authors who recommend that the term adenoma, which usually carries a benign connotation, be abandoned in favor of a descriptive diagnostic term, such as "BE-associated polypoid dysplasia." BE patients with this lesion should be considered strong candidates for esophageal resection similar to lesions of this kind that occur in inflammatory bowel disease.