Abstract: Objective To analyze the pathological characteristics of superficial infiltrating squamous cell carcinoma of endoscopic submucosal dissection (ESD). Methods 187 cases of invasive squamous cell carcinoma after ESD operation were collected from 2016 Jan 31 to 2017 Dec 31. The tumor differentiation, invasion depth, infiltrative growth pattern (INF), tumor budding, angiovascular lymphatic invasion and margin were determined. The pathological diagnosis of endoscopic biopsy and surgical operation after ESD were searched. Results The patients were aged from 42 to 83 years old, including 147 males and 40 females. 9.1% patients had carcinoma/intraepithelial neoplasia in other sites, among which gastric adenocarcinoma was the most common one. Well, moderately and poorly differentiated squamous cell carcinoma accounted for 0.5%, 41.7% and 15.0%, respectively, while the remaining 42.8% cases were microinvasion and were difficult to be graded. Mucosa lamina propria, muscularis mucosa and submucosa invasion accounted for 39.6%, 32.6% and 27.8%, respectively. Submucosa infiltration ＜200 μm (SM1) accounted for 9.1% and submucosa infiltration ≥200 μm (SM2) accounted for 18.7%. Lymphatic vessel invasion was related to the depth of tumor invasion, tumor budding, INF. The invasion rate of lymphatic vessels increased with the increase of infiltration depth and the grade of tumor budding. The lymphatic invasion rate in INFb/c group was higher than that in INFa group. There was no statistical difference in the incidence of lymphatic vessel invasion between well/moderately differentiated squamous cell carcinoma and poorly differentiated squamous cell carcinoma. Multivariate analysis showed that tumor budding was an independent risk factor for lymphatic vessel invasion. The complete resection occupied 69.5% (130 cases), while most ofincomplete resection cases (57 cases) were involved by low-grade intraepithelial neoplasia. 69 cases had biopsy after ESD, among which there were 46 cases (66.67%) with no recurrence, 19 cases (27.54%) with recurrence, and 4 cases (5.80%) occurring in other sites. There was no statistical difference in recurrent rate between the complete resection (28.3%, 13/46) and the incomplete resection (31.6%, 6/19, P>0.05). Conclusion Tumor invasion depth, INF, tumor budding andlymphatic vessel invasion should all be disclosed for the ESD specimen pathological report. Tumor budding was an independent risk factor for lymphatic vessel invasion.