An 81-year-old woman presented with a complaint of mild epigastralgia. Esophagogastroduodenoscopy (EGD) revealed a reddish elevated lesion developing like a submucosal tumor, with a shallow depression approximately 20 mm in size, located in the second portion of the duodenum, just proximal to the papilla of Vater (Fig. 1a, b). EGD with magnified narrow-band imaging revealed irregular villous or papillary structures on the surface of the lesion (Fig. 1c). On endoscopic ultrasound, a coarse echoic mass extended from the mucosal layer to the shallow submucosal layer (Fig. 1d, arrow). A large number of Brunner's gland (BG) hyperplasia were observed in the duodenal bulb and superior duodenal angulus. Based on endoscopic and biopsy findings, the diagnosis of duodenal adenocarcinoma with submucosal invasion was made.

Endoscopic image showing a reddish elevated lesion, which developed like a submucosal tumor, in the second portion of the duodenum (a, b). Magnifying endoscopic findings using narrow-band imaging showing irregular villous or papillary structures on the surface of the lesion (c). Endoscopic ultrasound showing a coarse echoic mass in the mucosal to the shallow submucosal layer (d, arrow). [Color figure can be viewed at wileyonlinelibrary.com]

Considering the patient's health status, we proceeded with a local wedge resection. Pathological examination of the resected specimen confirmed the presence of tumors with a tubular or papillary structure, consisting of cuboidal and columnar cells with lobule-like glands (Fig. 2a), having spread throughout the mucosa including partial invasion of the submucosa with apparent desmoplastic stromal reactions (Fig. 2a, arrow; and Fig. 2b). The proliferating components resembled BGs. On immunohistochemistry, the tumor cells were positive for MUC5AC (Fig. 2c) in the upper layer and MUC6 (Fig. 2d) in the deeper layer and negative for MUC2, CD10, and CDX2. The final diagnosis was confirmed as a gastric-type submucosal invasive adenocarcinoma.

Pathological examinations of the resected specimen showing tumors with a tubular or papillary structure consisting of cuboidal and columnar cells with lobule-like glands (a) (hematoxylin and eosin [HE], 4×), spreading throughout the mucosa and partially invading the submucosa, with apparent desmoplastic stromal reactions (a, arrow; and b [HE, 100×]). Immunohistochemical examination showing that the tumor cells are positive for MUC5AC (c) in the upper layer and MUC6 (d) in the deeper layer. Image showing localization of Ki-67-positive cells in the superficial layer (e). The MIB-1 index in the infiltrated area is 5% (f). [Color figure can be viewed at wileyonlinelibrary.com]

Non-ampullary duodenal adenocarcinomas are uncommon neoplasms classified into two major subtypes, intestinal and gastric, according to their mucin phenotype. Gastric-type adenocarcinomas of the duodenum (GAD) arise more frequently in the first and second portions of the duodenum. Macroscopically, GADs tend to present as an elevated lesion but can also present as a submucosal tumor-like mass, as in our case. An oval-shaped marginal epithelium and dilatation of the intervening part, which reflect villous or papillary structures as in our case, are characteristic of GADs.

GADs often grow in the background of gastric foveolar metaplasia (GFM), heterotopic gastric mucosa, and BG hyperplasia. There is a correlation between GFM occurrence and BG proliferation towards the mucosal surface of the duodenum upon induction of GFM in the duodenal mucosal repair process. Therefore, GFM could be a precursor to gastric-type duodenal carcinogenesis, as in our case. As such, BG regeneration/proliferation might be associated with the pathogenesis of GAD. Of note, our case clearly demonstrates the very early process of submucosal invasion of GAD, with apparent desmoplastic stromal reactions. The evaluation of the biological behavior of GAD has not yet been established. Some reports have described that GADs may invade the submucosal layer, even if they are small in size and that GADs are associated with poorer overall survival than intestinal-type adenocarcinomas of the duodenum. However, in our case, Ki-67-positive cells tended to be localized in the superficial layer (Fig. 2e), with a calculated MIB-1 index in the infiltrated area of 5% (Fig. 2f). Furthermore, there was no overexpression of p53 protein, suggesting that the biological behavior was not high, despite infiltrative growth with desmoplastic reaction.

The method of treatment of duodenal adenocarcinomas should be carefully determined after considering their mucin phenotype, rather than simply opting for endoscopic treatment without further consideration.