Endoscopic mucosal resection (EMR) is increasingly used for management of Barrett esophagus (BE)-related neoplasia. Duplication of the muscularis mucosae (MM) has been described in BE esophagectomy specimens, where it can pose difficulties with accurate staging of carcinoma. The frequency, morphologic characteristics, and effect of MM duplication in adenocarcinoma staging in EMRs have not yet been evaluated. We studied 122 EMR specimens from 100 patients from 1999 to 2006. The following histologic features were scored: depth of EMR, presence of MM duplication and its extent, prolapse changes (extension of smooth muscle into lamina propria), gland entrapment, and diagnosis (original and study/final). Carcinomas reaching the level of submucosa were classified as invasive adenocarcinoma (INV); those confined to lamina propria or MM were classified as intramucosal adenocarcinoma (IMAC). Of 122 EMRs, 11 (9%) reached mucosa only, 109 (89%) extended to submucosa, and 2 (2%) extended into muscularis propria. MM duplication was present in 67% (75 of 111 specimens that reached at least submucosa). Prolapse changes were noted in 65 (54%) cases and gland entrapment in 67 (56%). Final pathologic diagnoses were 9 (7%) no specialized Barrett mucosa, 4 (3%) BE without dysplasia, 13 (11%) low-grade dysplasia, 51 (42%) high-grade dysplasia, 33 (27%) IMAC, and 12 (10%) INV. EMRs without BE were less likely to show MM duplication (P = 0.01) and there was a trend toward less frequent prolapse change (P = 0.08) and less gland entrapment (P = 0.08) as compared with EMRs with BE. However, there were no significant differences with respect to MM duplication, prolapse change, or gland entrapment between BE with or without dysplasia, IMAC, or INV. Among 33 cases of IMAC, tumor invaded lamina propria in 10 (30%), inner or single MM in 14 (42%), space between duplicated MM in 5 (15%), and outer MM layer in 4(12%). Lymphatic invasion was seen in 2 (10%) cases in which tumor reached the space between MM layers. Overstaging of carcinomas occurred in the original reports in 8 (7%) cases due to misinterpretation of the muscular anatomy, including one case in which the deep MM was interpreted as muscularis propria. These results show that MM duplication is commonly seen in EMR specimens. It is closely associated with the presence of BE but is not affected by neoplastic progression in the Barrett epithelium. Pathologists need to be aware of this distinctive anatomy of BE for accurate staging of adenocarcinomas, particularly to avoid mistaking a thickened outer MM as muscularis propria. Level of IMAC may be a critical feature because of potential access to lymphatic spaces between duplicated MM layers, and we therefore recommend including an explicit statement about the depth of adenocarcinoma invasion rather than using only broad terms such as IMAC or INV in the diagnostic report.