ObjectiveTo compare radical surgery with definitive chemoradiation (CRT) for esophageal squamous cell carcinoma using propensity score (PS) matching at our single institution.Materials and methodsA total of 386 consecutive, surgically treated and 243 CRT-treated cases between 2001 and 2014 were analyzed. PS was calculated using multivariable analysis (logistic regression) for pairs of variables such as treatment time, age, sex, primary tumor location, clinical stage, and clinical T- and N-stage for patients after excluding clinical T4 and M1 cases. According to PS, 133 surgically-treated and 134 CRT-treated cases were selected randomly by software.ResultsThe patients’ median age was 68 years in the CRT group and 71 years in the surgery group. Clinical stage II-III, T3, N0 (according to the 7th American Joint Committee on Cancer-2009), and upper plus middle thoracic esophageal disease were seen in 68%, 44%, 54%, and 59%, respectively, in the CRT group and 64%, 47%, 55%, and 64%, respectively, in the surgery group. The 3- and 5-year overall survival was 47.1% and 34.0% in the CRT group and 68.3% and 54.4% in the surgery group (p = 0.0019). The 3- and 5-year progression-free survival was 45.3% and 38.8% in the CRT group and 61.1% and 54.4% in the surgery group (p = 0.022).ConclusionCRT may be inferior to surgery in survival, although a selection bias for patients selected for a non-operative approach cannot be excluded, especially since surgery is the standard of care at this institution. A prospective randomized clinical trial will be necessary to draw a definite conclusion.