A myoepithelial hamartoma of the stomach is a rare tumor. It is also referred to as myoglandular hamartoma, adenomyomatous hamartoma, gastric adenomyoma and heterotopia [5, 6]. Lesions where pancreatic tissue is normally arranged are defined as an ectopic pancreas ; one third of patients may present with nests of pancreatic tissue dispersed in the smooth muscle bands. The exact pathogenesis and natural history of these lesions is not fully understood, possibly because very few cases have been reported in the literature. Generally, these lesions are either considered a form of hamartoma or a pancreatic heterotopia. Gal et al. suggested that these lesions should be considered a form of hamartoma based on their observation of goblet cells, argentaffin cells and smooth muscle stroma. Alternatively, Takahashi and Fukusato  reported that these lesions demonstrated presence of cytokeratin 7 on immunohistochemical staining, which is distributed in the pancreatic duct epithelium, suggesting that the heterotopic pancreas theory may be more plausible. A properly structured follow-up study may explicate the natural history of myoepithelial hamartoma.
According to our research, 39 cases have been reported so far in the literature, not including ours . In these cases, the patients’ age ranged from eight weeks to 81 years and in almost two thirds of them, the lesion was detected between the fourth and sixth decades of life . The distinctive histological feature of this tumor was the presence of hypertrophic smooth muscle bands that surround the epithelial elements, arranged in various patterns including simple glandular, Brunner’s gland, and pancreatic acini. Ectopic pancreatic tissue nests were reported in one third of the cases [5, 11] and were generally exocrine in nature, whereas the remainder had dilated pancreatic-like nests surrounded by smooth muscle bands, as in our case.
Lesions like these have been recorded in the gastric antrum (85%) and in the pylorus (15%) . They usually present with nonspecific symptoms such as intermittent gastric pain, nausea and vomiting, and in some patients cause intermittent pyloric obstruction [4, 5, 11]. Among the cases reported in the literature, most were diagnosed by using conventional X-rays of the upper gastrointestinal tract using a barium meal follow-through; others were detected incidentally during surgery or autopsy [4, 5, 11]. The diagnosis had not been suspected prior to postoperative or postmortem histological findings, even in the cases where an esophagogastroduodenoscopy had been performed [4, 5, 11]. Because these tumors exhibit the form of a submucosal mass with central pancreatic heterotopia, an endoscopy does not permit a differential diagnosis between hamartoma and leiomyoma, melanoma, lymphoma, carcinoid tumor, Kaposi’s sarcoma or eosinophilic granuloma . These hamartomas often may appear as a roundish- or oval-shaped mass in the submucosal tissue, simulating a lipoma, a neurofibroma or a polypoid formation.
Our case simulated a GIST, presenting on CT as an oval-shaped mass in his upper abdomen, compressing his duodenum. Endoscopy does not produce any useful biopsy for histology because most of these lesions, including the one in our case, are localized in the submucosal layer. Vandelli et al. had attempted to treat an antral lesion unsuccessfully but had made a correct preoperative diagnosis of hamartoma as they had removed it piecemeal via endoscopy. However, this approach failed and confirmed the difficulty of preoperative diagnosis. These observations, along with the negative histology results, corroborate the recommendation made by Hedembro et al. for surgery and resection in all cases of suspected gastric lesions to ensure a correct diagnosis. Radical surgery is justified in cases of a definite diagnosis of hamartoma as the lesion is very rare and may not be benign, and the risk of cancerization is still an open question . In case of an adenomyoma of the periampullary region, surgical limited resection is recommended; however, pancreaticoduodenectomy is often performed as the lesion is frequently misdiagnosed as a carcinoma . Whenever possible, perioperative frozen section diagnosis should be used to prevent unnecessary radical surgery.