Background: Chemoradiotherapy (CRT) followed by surgical resection is the standard of care for treating advanced esophageal cancer. However, the role of surgery has come into question in recent studies. The purpose of this study is to compare outcomes of patients treated with CRT with or without surgery. Methods: An IRB-approved database was queried to identify esophageal cancer patients treated with CRT with or without surgical resection between 2000 and 2011. Overall survival (OS) and disease-free survival (DFS) were calculated by the Kaplan-Meier method and log-rank analysis. Multivariate analysis for OS and DFS were calculated with a Cox proportional hazard ratio model. Results: We identified 232 patients treated with CRT (122 without surgery, 110 with surgery). Surgery was associated with a significant increase in OS and DFS. Median and 5 year OS for surgical versus nonsurgical patients was 42.2 months, and 42.3% versus 20.4 months and 29%, respectively (p = 0.0003). Median and 5 year DFS for surgical versus nonsurgical patients was 16.8 months and 29% versus 8.4 months and 22.8% (p < 0.001). MVA for OS revealed that lower stage (p = 0.0098), tumor length <5 cm (p = 0.0059), and surgery (p<0.0001) were prognostic for significantly decreased mortality, while age, gender, histology, tumor location, radiation dose, and radiation technique were not prognostic. MVA for DFS showed that tumor length <5 cm (p = 0.0112), radiation technique (p = 0.0023), and surgery (p = 0.0007) were prognostic for significantly decrease mortality, while lower stage (p = 0.069) and squamous histology (p = 0.055) were trending for decreased mortality. Age, gender, radiation dose, and tumor location were not prognostic for DFS. Conclusions: Surgery after CRT is strongly associated with increased OS and DFS in our esophageal cancer patient population. While we highly recommend surgical resection as part of trimodality treatment, it should only be performed in high volume centers. Longer followup in the already conducted randomized trials involving squamous cell carcinomas are needed to better qualify the initial negative results and randomized trials are need to address the role of surgery for adenocarcinomas.