Endoscopic mucosal resection of gastric neoplasms is a curative technique that avoids surgery and its potential complications. Infrequently performed in the West, the limitations, pitfalls and challenges provided by this new therapeutic modality are not well known by general surgical pathologists. We evaluated a series of 39 endoscopic mucosal resections and assessed the correlation between original biopsies and final diagnoses, depth of excision, status of deep and lateral margins, artifactual changes and recurrence rate. The tumors consisted of 24 intramucosal carcinomas, six high-grade dysplasias, eight low-grade dysplasias and one submucosal invasive carcinoma. The preresection diagnoses corresponded to the final evaluation in 63% of the cases with previous biopsies. In 37% of the cases, the biopsies under-diagnosed the neoplasia. The rate of positive margins was 38%. Iatrogenic changes, that is, intramucosal hemorrhage and electrodiathermic burn, were noted in 44% of the cases but hindered the pathologic evaluation in only 10% of the cases. Persistence or recurrence was observed in only seven cases and there was no progression to advanced adenocarcinoma. Based on our experience, we offer some recommendations in order to provide optimal pathologic analysis of endoscopic mucosal resection specimens.