Background Esophageal carcinoma is an aggressive malignancy and long-term survival is poor. Endoscopic ultrasound (EUS) is an additional staging modality to assess locoregional extent of this disease. We hypothesized that EUS may improve survival through more effective staging and better optimization of treatment. Methods We performed a retrospective review of all patients presenting with esophageal cancer at our institution from 1993 to 2003 ( n = 97) and compared outcomes between patients who underwent staging EUS and computed tomography (CT) versus CT alone. Survival was calculated using Kaplan-Meier methods and compared between groups using the log-rank test. Mean survival was compared using analysis of variance (ANOVA) methods. Results Overall 3-, 6-, and 12-month survival did not differ between the 2 groups (EUS: 92%, 84%, and 80% and CT: 83%, 67%, and 43%, log-rank P = .1), which held true despite stratification by treatment modality (all P >.1). The mean survival for the EUS group was 16 ± 3 months and for the CT group, 12 ± 1.5 months ( P = .2). Further analysis by stage showed no difference in survival between the 2 groups (all P >.1). However, stage 2A and 3 surgical patients had better survival than nonsurgical patients (both P = .02) irrespective of staging modality. EUS patients were no more likely to receive surgical, neoadjuvant, or definitive chemoradiation than CT patients (all P >.1). Conclusions Overall survival as well as survival by stage did not differ between patients who underwent staging via EUS and CT versus CT alone, and patients staged with EUS were not more likely to receive any one intervention. Irrespective of staging modality, stage 2A and 3 patients who underwent surgical intervention had better survival than those who did not receive an operation.