This study aimed to identify clinicopathological factors predictive of lymph node metastasis in patients with the poorly differentiated early gastric cancer (EGC) to assess the feasibility of using endoscopic submucosal dissection (ESD).

The records of patients with poorly differentiated early gastric cancer undergoing gastric radical resection between January 2012 and December 2016 were reviewed in Ruijin hospital. Those with distant metastasis, two or more malignant tumors, remnant gastric cancer, neo adjuvant therapy, previous history of gastric surgery or clear history of perigastric lymphadenectomy, and mixed tumors were excluded. Age, sex, presence of ulcerous lesion, tumor size, tumor location, depth of invasion, type of differentiation, lymphatic vessel invasion, vascular invasion, nerve invasion and HER2 expression were collected. Univariate and multivariate stepwise logistic regression analyses were used to identify the independent risk factors of perigastric lymph node metastasis.According to the Guidelines for the Treatment of Gastric Cancer (2018 edition) of the Chinese Society of Clinical Oncology (CSCO), the expanded indications of ESD for EGC are as follows: (1)no ulcerative lesions, the maximum diameter of lesions >2 cm of differentiated intramucosal cancer; (2)ulcerative lesions, the maximum diameter of lesions ≤3 cm of differentiated intramucosal cancer; (3)no ulcerative lesions, undifferentiated intramucosal carcinoma with diameter ≤2 cm. The relationship between clinicopathological factors and lymph node metastasis was analyzed.

A total of 517 patients, aged 21-83 (57.1±11.7), including 307 males and 210 females, were enrolled in the study. Among them, 114 (22.0%) patients had lymph node metastasis. Univariate analysis showed that ulcerative lesion (P=0.042), tumor diameter (P=0.048), depth of invasion (P<0.001), location of tumors (P<0.001), lymphatic vessel invasion (P=0.009), vascular invasion (P<0.001) and nerve invasion (P=0.028) were related to lymph node metastasis after radical resection of poorly differentiated early adenocarcinoma. Age, sex, type of differentiation and HER2 expression were not significantly correlated to lymph node metastasis (P>0.05). Multivariate analysis showed that tumor size (OR=1.61, 95% CI: 1.03-2.52, P=0.037), depth of invasion (OR=2.77, 95% CI:1.66-4.63, P<0.001), lymphatic duct invasion (OR=14.74, 95% CI:1.58-137.36, P=0.018) were independent risk factors for lymph node metastasis in poorly differentiated EGC, and ulcerative lesion was not a risk factor for lymph node metastasis (OR=0.82, 95% CI:0.56-1.18,P=0.285). A total of 119 patients with poorly differentiated EGC fully complied with the relative indications of ESD recommended by the Japanese Statute and the criteria for radical resection after ESD. Among them, 14 (11.8%) still had perigastric lymph node metastasis, while the gender, tumor diameter, location, differentiation and HER2 expression were not associated with lymph node metastasis (P>0.05).

For patients with poorly differentiated EGC, the application of ESD should be carefully weighed with precise assessment of tumor diameter, depth of invasion, and lymphatic duct invasion.

Stomach neoplasms,early; Stomach neoplasms,poorly differentiated; Endoscopic submucosal dissection; Lymph node metastasis