Background: Most studies of pre-treatment positron emission tomographic (PET-CT) scanning have focused upon its use in tumor staging. The role of PET-CT in predicting survival has received comparably less attention. As such we sought to assess the relationship of pretreatment SUVmax to overall survival in surgical patients with esophageal cancer. Methods: The study population consisted of 57 patients who underwent esophagectomy for adenocarcinoma (n=47) or squamous cell carcinoma (n=10) of the esophagus. Patients were taken from a cohort of 116 consecutive esophagectomy patients between 7/2005 and 9/2008. PET-CT scanning was performed at a single center and SUVmax was recorded for each patient prior to any therapy. Ten of the 57 patients received neo-adjuvant chemotherapy (n=7) or chemoradiation (n=3). Median follow-up was 14.5 months. Survival was assessed using Kaplan-Meier analysis. Results: The median SUVmax on pretreatment PET-CT was 6.4 (range 0-59). Receiver operating characteristic (ROC) curve identified a SUVmax of 4.6 to optimally discriminate survival. Patients with low SUVmax (<4.6; N=21) had significantly (p<0.001) better survival (median survival not reached) than those with high SUVmax (>4.6; N=36; 19.27 months). Interestingly, Stage 3 patients with low SUVmax (n=8) had significantly better survival (p<0.05) than Stage 3 patients with high SUVmax (N=17). Likewise, N1 disease patients with low SUVmax (N=11) had significantly better survival (p<0.05) than high SUVmax N1 patients (N=26). There was a significant association between SUVmax and death on univariate logistic regression (p=0.022, HR 1.03, CI 1.00-1.06); age, gender, endoscopic tumor length, and pN, were not significant. When controlling for stage and grade on multivariate analysis, SUVmax remained independent of grade (p<0.05), but not stage (p=0.093). Conclusions: Pretreatment PET-CT SUVmax is strongly associated with survival in patients with esophageal carcinoma. Longer overall survival of Stage 3 or N1 patients with low pretreatment SUVmax may suggest more favorable tumor biology in this cohort of patients. This may be valuable prognostic information for patients with advanced disease.