Objective: The objective was to identify whether repeat positron emission tomography scan after neoadjuvant chemoradiotherapy in patients with esophageal cancer predicted a complete response. Methods: A retrospective study using a prospective database was performed. Patients had esophageal cancer and underwent neoadjuvant chemoradiotherapy, an initial and repeat positron emission tomography, endoscopic ultrasound with fine-needle aspiration (at the same institution), and Ivor Lewis esophagogastrectomy with lymph node resection. Results: There were 221 patients who underwent Ivor Lewis, 86 of whom had their initial and repeat positron emission tomography scans performed at the same center. Of these, 37 patients (43%) were complete responders. The median maximum standardized uptake value of esophageal cancer decreased by 72% in the 37 patients who were complete responders, by 58% in the 31 patients who were partial responders, and by 37% in the 18 patients who had a minimal pathologic response. When the maximum standardized uptake value decreased by more than 64%, the patient was likely to be a complete responder (P = .003, area under the curve = 0.75). Conclusion: When initial and repeat positron emission tomography scans are performed at the same center at least 30 days after the completion of preoperative chemoradiotherapy, the percent change in the maximum standardized uptake value is a predictor of the response to chemoradiotherapy by a patient with esophageal cancer. When the maximum standardized uptake value decreases by 64% or more, it is likely that the patient is a complete responder. These data may help guide neoadjuvant therapy and identify patients for a future randomized study that compares observation with surgical resection in patients with esophageal cancer who appear to be complete responders. © 2009 The American Association for Thoracic Surgery.