Background Positron emission tomography scanning is used to assess response to induction therapy after treatment of esophageal cancer. A decrease in standardized uptake value has been correlated with response to therapy, with a standardized uptake value of zero often assumed to indicate complete absence of disease. We hypothesize that a significant number of patients may have residual esophageal cancer despite complete metabolic response (CMR). Methods A prospective database was reviewed for esophageal cancer patients receiving induction therapy followed by esophagectomy on whom both preinduction and postinduction positron emission tomography scans were obtained. Patients with a postinduction SUV of 0 (or "no uptake") were categorized as complete metabolic responders. Survival was calculated by the Kaplan-Meier statistic. Results Among 120 patients, 32 (27%) had postinduction CMR after chemotherapy (21 of 81, 26%) or chemoradiation (11 of 39, 28%). At surgery, 19 patients (59%) with CMR had residual disease, including 12 (38%) with nodal metastases. Even among patients with a negative postinduction biopsy, 4 of 10 (40%) had residual disease. Final pathologic stages of patients with CMR were yp0 (complete pathologic response) in 13 (41%), ypI in 4 (12%), ypII in 9 (28%), and ypIII in 6 (19%). Three-year survival was 83% in the CMR group versus 41% in the remainder of the cohort (p = 0.02). Conclusions A CMR on postinduction positron emission tomography scan predicts but should not be assumed to be synonymous with complete pathologic response in esophageal cancer patients. The presence of residual disease should be strongly considered despite CMR and negative biopsy in patients receiving induction therapy.