Introduction: For clinical assessment of neoadjuvant radiochemotherapy of esophageal cancer reliable in-vivo methods are necessary. Therefore, the capabilities of F-18-Fluorodesoxyglucose-PET in comparison to histomorphological grading of tumor regression were studied. Methods: In 33 patients with locally advanced esophageal carcinoma (uT3, uN0-1, cM0) F-18-FDG-PET was performed before and 2weeks after radiochemotherapy. All tumors were resected by transthoracic en-bloc esophagectomy 3-4 weeks after induction therapy. A subgroup of 11 patients underwent weekly PET scan during neoadjuvant therapy.PET was performed in a dedicated scanner 1.3 h after administration of 370 MBq F-18-FDG. Data analysis based on maximum SUV data derived from individual regions of interest in pre- and posttherapeutic images. PET data were compared to histomorphological grading parameters for tumor regression whithin the resected tissues. Results: The comparison of histopathological tumor regression after neoadjuvant therapy and PET SUV differences showed a significant X2 P-value of 0.006. There was a significant decrease of the SUV data from 9.1±3.5 to 4.3±1.9 (P＜0.0001). In therapy responders SUV was diminished by 59% and in non-responders by 34 %. Longitudinal SUV measurement during neoadjuvant therapy showed a strong SUV decrease already after one and two weeks (P=-0.021 and 0.003). Conclusion: The recent data of the FDG-PET follow-up after neoadjuvant therapy show that PET is able to predict therapy response.Longitudinal PET data advocate that it may be possible to recognize response also very early during radiochemotherapy.