Objectives: Esophageal squamous cell carcinoma (ESCC) is associated with abnormal expression of multiple genes. Neuroepithelial transforming gene 1 (NET1) is a carcinogenic gene, overexpressed in several cancers. The tumor suppressor gene CD82, which encodes the protein KAI1, is down regulated in cancers. Lymphatic vessel invasion (LVI) is closely associated with metastasis. We investigated the expression levels of NET1, KAI1/CD82, LVI and their correlation with clinicopathological factors in ESCC. Methods: Immunohistochemistry and Western Blot were used to detect NET1, KAI1/CD82 expression levels in the para-carcinoma tissue and ESCC. LVI detected by D2-40 immunohistochemical staining. The relationships between the protein expression levels, positive LVI and clinicopathologic data were analyzed. Eighty-five patients, who had a primary resection of esophageal cancer, were analyzed by univariate and multivariate logistic regression, and univariate and multivariate survival analysis. Results: NET1 expression levels in cervical cancer were significantly higher than in the para-carcinoma tissue, (P<0.05). KAI1/CD82 expression was markedly lower in ESCC than in the para-carcinoma tissue (P<0.05). NET1 high expression and positive LVI were positively correlated while KAI1/CD82 expression was negatively correlated with invasion, lymph node metastasis, and clinical stage. They were all independent prognostic factors for lymph node metastasis. Kaplan-Meier analysis revealed that NET1 high expression and positive LVI was negatively correlated with overall survival (OS), while KAI1/CD82 expression was positively correlated with OS. Low KAI1/CD82 expression, high expression of NET1, positive LVI was associated with a poor prognosis in ESCC. Multivariate Cox regression analysis indicated that the positive LVI were independent predictors for OS in ESCC. Conclusions: NET1 high expression, positive LVI and KAI1/CD82 negative expression were all independent prognostic factors for lymph node metastasis in ESCC. Positive LVI was also an independent prognostic factor for OS. Combined detection of these factors may be of significant value in predicting the prognosis and metastasis in ESCC patients.