In summary, we report a gastric fundic mass consisting of readily distinguishable epithelial and mesenchymal components. The epithelial cells showed morphology and immunophenotype consistent with metaplasia and downgrowth of gastric glands into the submucosa, suggesting GCP. The mesenchymal component was comprised of ovoid-shaped cells harboring EWSR1-CREM fusion. Given the absence of lymph node, vascular or neural invasion, low mitotic activity, and disease-free status at 28-month follow-up, the GCP and mesenchymal tumor were likely benign.
A panel of mesenchymal tumors of the GI tract was considered for the mesenchymal component. CCSLGT was considered based on the anatomic site and presence of EWSR1-CREM fusion. However, CCSLGT is a malignant entity that occurs mostly in young adults. In a series of 3 CCSLGT cases that harbored EWSR1-CREB family fusion, the patient age ranged 24-29, and one developed lung metastasis and died after 2 years (11). Also, all three patients displayed patchy positivity for S-100 and negativity for CD117 on immunohistochemistry, further excluding a CCSLGT diagnosis (12). The immunostaining results in this case agreed with gastrointestinal stromal tumor (GIST) on CD117 positivity and S-100 negativity. However, CD117 negativity occurs in up to 18% of gastric GISTs, making it a less specific marker (12). Furthermore, DOG-1 immunoreactivity in reported in 87%-97% GISTs (13) and CD34 in 60-70% (12), both of which were lacking in this case. Morphologically, most GISTs are composed of either short fascicles of spindle cells or round, epithelioid cells arranged in nests, which contrasts with the ovoid cells arranged in sheets in this case. Other tumor types, such as smooth muscle neoplasms and tumors with fibroblastic/myofibroblastic, neural, or melanocytic differentiation, were excluded by respective negative immunostaining for SMA, desmin, neuroendocrine markers, and melanocytic markers in this case (12). Recent years have also seen an explosion in the histological spectrum of mesenchymal tumors with EWSR1/FUS-CREB family fusions (2–4, 14). Shibyama et al. reported 8 cytokeratin-positive intra-abdominal malignancies harboring EWSR1/FUS-CREB fusions (14). Among these cases, one was strikingly resemblant to ours in that a gastric submucosal mesenchymal neoplasm was admixed with dilated epithelial structures that invaginated from the mucosa. EWSR1-CREB fusion was detected with fluorescence in situ hybridization (FISH), and the tumor also showed low mitotic activity. There were, however, peripheral lymphoid cuffing and prominent hemorrhagic pseudoangiomatous spaces suggested of AFH. In addition to cytokeratin positivity, the tumor also showed other discrepancies such as epithelioid cytomorphology, diffuse SMA positivity, and possibly a more aggressive clinical course. Interestingly, 2 subsequent specimens resected in the abdominal cavity at recurrence displayed similar histology but without the epithelial inclusions, suggesting the GCP-like structure as a completely independent or stomach-specific phenomenon. However, conclusive elucidation awaits further research.
In addition to the case series by Shibyama and colleagues, Argani et al. also described 13 EWSR1/FUS–CREB-rearranged malignant epithelioid neoplasms, most of which located intra-abdominally or expressing epithelial markers cytokeratin and/or EMA (4). An interesting finding was a EWSR1-CREM fusion that attaches CREM exon 7 to EWSR1 exon 15, the same composition as with this case. One of the EWSR1-CREM-rearranged tumors was intra-abdominal and inseparable from gastric fundus, although it was not specified which breakpoints were detected. Since EWSR1 exon 15 is a highly rare breakpoint, this similarity may reflect shared molecular basis between these cases and ours despite immunohistochemical and clinical differences. As new EWSR1/FUS–CREB-rearranged mesenchymal tumors are detected and new classes proposed at a rapid pace, it is possible that this case falls into one of the recorded classes, although more insights into the relationship between the pathogenesis of GCP and the mesenchymal neoplasm are warranted.
Gastritis cystica profunda was previously associated with a history of gastric surgery, although a recent review found GCP occurring approximately equal percentages of patients with (52%) and without surgery (48%) (6). However, GCP was observed in animals after gastrectomies or predisposition to H. pylori infection, suggesting association with insult to the gastric mucosa (6, 9). Highly suspected as precancerous, GCP was reported in 3 cases as the background from which early gastric adenocarcinoma arose (8–10). In this case, it is unknown whether GCP and mesenchymal neoplasm developed independently from one other or in concert, although considering the rarity of both entities and the precancerous nature of GCP, the two lesions may be pathogenetically connected.
To our knowledge, this is the first to report a EWSR1-CREM fusion in a gastric mesenchymal tumor with accompanying GCP. Whether this case suggests a novel entity or falls into one of proposed classes awaits report of more similar cases and insights into the relationship between GCP pathogenesis and oncogenesis.