Purpose: To determine the ability of 2-deoxy-2-[18F] -fluoro-D-glucose (FDG) positron emission tomography (PET) to monitor response in locally advanced gastroesophageal cancer (LAGEC). Additionally, optimal FDG-PET methods for response monitoring were selected. Procedures: Sequential dynamic FDG-PET scans were performed in 13 patients with LAGEC (T2-3N0-1M0-1a) treated with neoadjuvant cisplatin and gemcitabine plus granulocyte macrophage colony stimulating growth factor at a three weekly schedule. The standard FDG-PET method, nonlinear regression (NLR), was compared with computed tomography (CT), endoscopic-ultrasound (EUS), and histopathology as well as with 21 simplified analytical FDG-PET methods. Results: Five out of 12 operated tumors responded histopathologically with less than 10% residual tumorcells (42%). These had a higher decrease in FDG uptake compared with nonresponders (P = 0.008). Early (after two cycles) and late (after completed induction therapy) response evaluation showed a specificity of 86% and 100%, respectively, and a sensitivity of 100%. Both FDG-PET and EUS were superior to CT. From 21 methods analyzing FDG uptake, the quantitative Patlak analysis, the simplified kinetic method (SKM), and the semiquantitative standardized uptake value corrected for bodyweight (SUV-BW) seemed to correlate best with NLR. Conclusions: FDG-PET reliably predicted response in LAGEC. FDG-PET measurements using Patlak analysis or the more clinical applicable SKM and SUV-BW were acceptable alternatives to NLR.