AIM:To identify the predictive clinicopathological factors for lymph node metastasis (LNM) in poorly differentiated early gastric cancer (EGC) and to further expand the possibility of using endoscopic mucosal resection (EMR) for the treatment of poorly differentiated EGC. METHODS: Data were collected from 85 poorly- differentiated EGC patients who were surgically treated. Association between the clinicopathological factors and the presence of LNM was retrospectively analyzed by univariate and multivariate logistic regression analyses. RESULTS: Univariate analysis showed that tumor size (OR = 5.814, 95% CI = 1.050 - 32.172, P = 0.044), depth of invasion (OR = 10.763, 95% CI = 1.259 - 92.026, P = 0.030) and lymphatic vessel involvement (OR = 61.697, 95% CI = 2.144 - 175.485, P = 0.007) were the significant and independent risk factors for LNM. The LNM rate was 5.4%, 42.9% and 50%, respectively, in poorly differentiated EGC patients with one, two and three of the risk factors, respectively. No LNM was found in 25 patients without the three risk factors. Forty-four lymph nodes were found tohave metastasis, 29 (65.9%) and 15 (34.1%) of the lymph nodes involved were within N1 and beyond N1, respectively, in 12 patients with LNM. CONCLUSION: Endoscopic mucosal resection alone may be sufficient to treat poorly differentiated intramucosal EGC (≤ 2.0 cm in diameter) with no histologically-confirmed lymphatic vessel involvement. When lymphatic vessels are involved, lymph node dissection beyond limited (D1) dissection or D1+ lymph node dissection should be performed depending on the tumor location. AIM: To identify the predictive clinicopathological factors for lymph node metastasis (LNM) in poorly differentiated early gastric cancer (EGC) and to further expand the possibility of using endoscopic mucosal resection (EMR) for the treatment of poorly differentiated EGC. METHODS: Data were collected from 85 poorly-differentiated EGC patients who were surgically treated. Association between the clinicopathological factors and the presence of LNM was retrospectively analyzed by univariate and multivariate logistic regression analyzes. RESULTS: Univariate analysis showed that tumor size (OR = 5.814, 95% CI = 1.050-32.172, P = 0.044), depth of invasion (OR = 10.763, 95% CI = 1.259-92.026, P = 0.030) and lymphatic vessel involvement (OR = 61.697, 95% CI = 2.144-175.485, P = 0.007 ) were the significant and independent risk factors for LNM. The LNM rate was 5.4%, 42.9% and 50%, respectively, in poorly differentiated EGC patients with one, two and three of the risk factors, eachl No LNM was found in 25 patients without the three risk factors. Forty-four lymph nodes were found to metastasis, 29 (65.9%) and 15 (34.1%) of the lymph nodes involved were within N1 and beyond N1, respectively, in 12 patients with LNM. CONCLUSION: Endoscopic mucosal resection alone may be sufficient to treat poorly differentiated intramucosal EGC (≤ 2.0 cm in diameter) with no histologically-confirmed lymphatic vessel involvement. When lymphatic vessels are involved, lymph node dissection beyond limited ) dissection or D1 + lymph node dissection should be performed depending on the tumor location.