Background: NAC-S is a standard treatment for patients (pts) with stage II/III (UICC-TNM 5th ed.) ESCC in Japan. CRT is also optionally performed for pts who refused or who are not tolerable to surgery. However, no randomized clinical trial has been conducted comparing NAC-S with CRT. The aim of this study is to explore subgroups in which survival outcome of pts undergoing CRT was potentially equivalent to NAC-S. Methods: The data of two trials for cStage II/III ESCC, JCOG9907 (accrual period: 2000-2006) and JCOG9906 (2000-2002) were used. J9907 demonstrated that NAC-S (2 course of cisplatin and 5-fluorouracil [CF] followed by surgery) showed superior OS compared with surgery followed by postoperative chemotherapy. J9906 was a single arm trial which explored the efficacy and safety of CRT (CF plus 60 Gy). Both two trials were conducted under similar eligible criteria. Subgroup analyses between two trials (serum Albmin (Alb), cT, cN, c-stage, and tumor location) using Cox proportional hazard model were performed for pts assigned to receive NAC-S in J9907 and pts in J9906. Results: All eligible pts for NAC-S group (n=164) in J9907 and all eligible pts in J9906 (CRT group: n=73) were analyzed. Pts backgrounds were similar between two groups. OS for NAC-S group was better than that for CRT group (adjusted HR 1.73; 95% CI 1.19-2.51). Subgroup analysis using Cox proportional hazard model demonstrated that all subgroups in NAC-S group showed shorter OS compared with those in CRT group (see Table). Conclusions: OS of NAC-S was favored over that of CRT. The subgroup in which OS of CRT was equivalent to that of NAC-S was not detected. For further improvement of CRT for this population, single arm trial (JCOG0909) is ongoing to evaluate the efficacy and safety of CRT modifying chemotherapy regimen and total radiation dose.