Background: The aim of this study is to compare the long-term outcomes between patients who underwent endoscopic submucosal dissection (ESD) and those who underwent surgery for early gastric cancers (EGCs) turned out to be out-of-indication of ESD. Methods: We retrospectively reviewed database on gastric cancer patient cohort who underwent ESD or surgery from 2004 to 2014. Patients whose lesions revealed to be out-of-indication for endoscopic resection after ESD or surgery were included in the analysis after 1:1 propensity score matching for baseline clinicopathologic characteristics. Overall survival (OS) and gastric cancer recurrence rates were compared between the matched ESD and surgery groups. Results: After propensity score matching, a total of 193 pairs of patients were included, and 126 pairs were ESD with additional operation group and 67 pairs were ESD without additional operation group. OS (5-year OS rates, 89.5% vs. 89.9%; P=.105) and gastric cancer recurrence rates (5-year recurrence rates, 2.6% vs. 1.6%; P=.080) were not different between the ESD and the matched surgery group. In the matched subgroups, ESD with additional operation group had comparable OS and gastric cancer recurrence rates in comparison with matched surgery group. However, ESD without additional operation group had significantly shorter OS (5-year OS rates, 76.8% vs. 86.3%; P=.032) and higher gastric cancer recurrence rates (5-year recurrence rates, 11.9% vs. 0%; P=.002) than matched surgery group. ESD without additional operation was also significant risk factor for overall mortality and gastric cancer recurrence in multivariate analyses. Conclusions: Patients who did not undergo additional operation after ESD for EGCs turned out be the out-of-indication had worse long-term outcomes. Additional operation should be recommended, when EGC was identified to be the out-of-indication after ESD.