Purpose : The aim of this study was to determine whether $^{18}F$-fluorothymidine (FLT) PET is feasible for the early prediction of tumor response to induction chemotherapy followed by concurrent chemoradiotherapy in patients with esophageal cancer. Methods : This study was prospectively performed as a collateral study of "randomized phase II study of preoperative concurrent chemoradiotherapy with or without induction chemotherapy with S-1/oxaliplatin in patients with resectable esophageal cancer". $^{18}F$-FLT positron emission tomography (PET) images were obtained before and after two cycles of induction chemotherapy, and the percent change of maximum standardized uptake value (SUVmax) was calculated. All patients underwent esophagography, gastrofiberoscopy, endoscopic ultrasonography (EUS), computed tomography (CT) and $^{18}F$-fluorodeoxyglucose (FDG) PET at baseline and 3-4 weeks after completion of concurrent chemoradiotherapy. Final tumor response was determined by both clinical and pathologic tumor responses after surgery. Results : The 13 patients for induction chemotherapy group were enrolled until interim analysis. In a primary tumor visual analysis, the tumor detection rates of baseline $^{18}F$-FLT and $^{18}F$-FDG PET were 85% and 100%, respectively. The tumor uptakes on $^{18}F$-FLT PET were lower than those of $^{18}F$-FDG PET. Among nine patients who completed second $^{18}F$-FLT PET, eight patients were responders and one patient was a non-responder in the assessment of final tumor response. The percent change of SUVmax in responders ranged from 41.2% to 79.2% (median 57.1%), whereas it was 10.2% in one non-responder. Conclusion : The percent change of tumor uptake in $^{18}F$-FLT PET after induction chemotherapy might be feasible for early prediction of tumor response after induction chemotherapy and concurrent chemoradiotherapy in patients with esophageal cancer.