In our study, positive lymph node metastasis was found to influence OSR statistically. Lymphatic invasion was the most important risk factor for lymph node metastasis in the univariate analysis.
Thus far, many prognostic factors have been proposed. For lymph node metastasis, risk factors were angiolymphatic invasion,9 the tumor-associated neutrophil-to-lymphocyte ratio (TA-NLR), which was calculated as the average number of neutrophils (CD15-positive cells) divided by the average number of T lymphocytes (CD8-positive cells),4 and tumor size and tumor grade in rectal NETs.10 This study showed that lymph node metastasis was not associated with vascular invasion and tumor size but instead with lymphatic invasion. The Ki-67 index and the TA-NLR were independent prognostic factors for recurrence-free survival (RFS) and OS.4 In this study, positive lymph node metastasis was an important factor for the OSR in the univariate analysis. This may be influenced by many censored cases, and multivariate analysis may produce different results. Furthermore, for type 3 GNETs without distant metastases, radical surgery, serum carcinoembryonic antigen (CEA) level, immunohistochemistry marker CD56, and chromogranin A (CgA) are important for prognostic evaluation.11 In this study, unfortunately, we did not obtain enough data to determine whether prognosis was affected by angiolymphatic invasion, but we would like to consider these factors in further research.
GNENs are rare neoplasms derived from the enterochromaffin-like (ECL) cells of the gastric mucosa that play a role in regulating gastric acid production.11, 12 GNENs are more common in male patients and the average age of onset is 55 years. The incidence of proximal gastric fundus and cardia region accounted for 69.7% of cases. The specificities of biomarkers in gastroenteropancreatic NET patients were 86% for CgA, 100% for neuron-specific enolase (NSE), 91% for CEA, and 100% for 5-hydroxyindoleacetic acid (5-HIAA). The corresponding sensitivities were 68% for CgA, 33% for NSE, 15.4% for CEA, and 35% for 5-HIAA.11 In p53 expression, it was negative in all NETs G3 but positive in almost 78% of poorly differentiated NENs.13 The proportions of tumor recurrence of the patients with GNEN were 72% of liver metastasis, 28% of peritoneal metastasis, and 25% of lymph node metastasis. On the other hand, the spleen, kidney, and brain are relatively rare sites of recurrence.4 Rindi et al. described three clinicopathologic subtypes of well-differentiated ECL cell tumors of the stomach: Type 1 was associated with body-fundus chronic atrophic gastritis, Helicobacter pylori-related gastritis, autoimmune gastritis, pernicious anemia, and hypergastrinemia; Type 2 was associated with hypertrophic gastropathy and hypergastrinemia due to multiple endocrine neoplasia type 1 (MEN-1) and Zollinger-Ellison syndrome; Type 3 was sporadic.3,14,15 ENETS reported that the overall metastatic risk is low in type 1 GNETs and has been directly correlated with tumor size (10 mm is the cut-off.), and recommended tumors > 10 mm were resected. The RFS rate of patients with type 1 GNETs for approximately 24 months is 100%. For type 2 GNETs, local resection can be recommended with or without duodenal or pancreatic NENs as part of MEN-1. For type 3 GNETs, while endoscopic management for small lesions has been proposed, surgical treatment such as gastrectomy with lymph node dissection following the strategy for gastric adenocarcinomas is recommended.8 GI carcinoids originate as hyperplastic proliferations of endocrine cells in the basal portion of the glands, develop into extraglandular buddings, and penetrate the muscularis mucosae, resulting in the formation of a submucosal nodule. Soga reported that GI carcinoids that penetrate the muscularis mucosae and enter the submucosal layer (SM carcinoids) are indicative of malignancy without exception.9 The previous study reported that 59 out of 381 cases (15.5%) of gastric SM carcinoids had metastases. The metastasis rates were 7.9% (carcinoids < 10 mm), 13.4% (> 10 mm and < 20 mm), and 26.7% (> 20 mm), respectively. Moreover, the average size of the SM G-carcinoids was 14.4 mm, and there was a significant difference between their size with metastases (20.9 mm) and without metastases (13.2 mm).9 In this study, there was no significant difference according to tumor size for OSR. This may be associated with insufficient data for the T category.
The 5-year survival rate in patients with gastric SM carcinoids was the same as that of patients with gastric SM carcinomas (89.6% versus 82.7%). The 5-year survival rate in patients with gastric SM carcinoids with metastases was 61.7%.9 Therefore, it is necessary to pay attention to the patients with early gastric SM carcinoid, especially those over 10 mm. In contrast, it has been reported that approximately 60% of patients with less than 20 mm type 1 NET had recurrence after endoscopic resection.2
In this study, two patients survived over 10 years. One patient survived for over 13 years despite recurrence due to the introduction of somatostatin analogs. In a previous study, 23% of the tumors regarded to be NET G3 were reported to be strongly somatostatin receptor 2A positive.13 The Japan Neuroendocrine Tumor Society recommends somatostatin analogs, everolimus, and streptozocin for GNETs and somatostatin analogs, etoposide plus cisplatin, irinotecan plus cisplatin, and etoposide plus carboplatin for GNECs as chemotherapy.16 In addition, Hainsworth et al. proposed a treatment with paclitaxel, carboplatin, and etoposide for advanced poorly differentiated neuroendocrine carcinoma because it produced an overall response rate of 53% and a complete response rate of 15%. On the other hand, they said it seems reasonable to treat patients with metastatic poorly differentiated NETs according to guidelines established for small-cell lung cancer.17 Furthermore, long-term prognosis can be expected because of the development of new drugs. Of the 14 cases of type 1 GNETs, one patient died due to liver metastasis 1 year after surgery. Because the tumor-related death rate of type 1 GNETs is 0%, we need more histopathological reconsiderations.
There are some limitations to this study. First, there are some missing data. Second, the diagnostic criteria were not standardized because of the retrospective nature of the study. Third, the small number of cases and short follow-up time may have influenced the statistically significant difference for OSR with or without lymph node metastasis. Finally, the differences in histological classification between Japan and the WHO have not been considered.
We evaluated the outcomes in patients with GNEN in our institute. Positive lymph node metastasis was found to influence OSR statistically. In addition, lymphatic invasion was the most important risk factor for lymph node metastasis in the univariate analysis. Further retrospective analyses, such as resected specimens with standardized criteria, can help us further our understanding of GNENs.