Objective To systematically review the prognostic significance of extranodal extension in gastric cancer, esophageal cancer and gastroesophageal junction cancer. Methods We searched EMbase, The Cochrane Library (Issue 9, 2016), PubMed, CBM, CNKI, VIP and WanFang Data databases from inception to August 2016, to collect cohort studies about the prognostic significance of extranodal extension. Two reviewers independently screened literature, extracted data and evaluated the risk of bias of included studies. Then, meta-analysis was performed by using RevMan 5.2 software. Results A total of 16 studies involving 9 445 participants were included. The results of meta-analysis showed that for gastric cancer patients, the overall survival (OS) (HR=1.71, 95%CI 1.49 to 2.00,P<0.000 01), disease free survival (DFS) (HR=1.39, 95%CI 1.12 to 1.73,P=0.003) and cancer specific survival (HR=1.52, 95%CI 1.18 to 1.96,P=0.001) in ENE(+) group were lower than ENE(–) group. For esophageal cancer and gastroesophageal junction cancer patients, the overall survival (OS) (HR=1.84, 95%CI 1.49 to 2.27,P<0.000 01), disease free survival (DFS) (HR=2.18, 95%CI 1.70 to 2.81,P<0.000 01) and cancer specific survival (HR=1.73, 95%CI 1.19 to 2.52,P=0.004) in ENE(+) group were lower than ENE(–) group. Conclusion Current evidence indicates that ENE(+) was correlated with a poor prognosis in gastric cancer, esophageal cancer and gastroesophageal junction cancer patients. Due to the quantity and quality limitations of included studies, more high quality cohort studies are needed to verify the above conclusions.