The optimal radiotherapy dose for non-operated localized esophageal squamous cell carcinoma (NOL-ESCC) patients undergoing concurrent chemoradiotherapy (CCRT) is hotly debated.

We identified eligible patients diagnosed within 2008-2013 from Taiwan Cancer Registry and constructed a propensity score matched cohort (1:1 for high dose (⩾60Gy) vs standard dose (50-50.4Gy)) to balance observable potential confounders. We compared the hazard ratio (HR) of death between standard and high radiotherapy dose groups during the entire follow-up period. We performed sensitivity analysis (SA) to evaluate the robustness of our finding regarding potential unobserved confounders & index date definition.

Our study population constituted 648 patients with well balance in observed co-variables. The HR of death when high dose was compared to standard dose was 0.75 (95% confidence interval 0.64-0.88). Our result was sensitive to potential unobserved confounders but robust to alternative index date definition in SA.

We found that higher than standard radiotherapy dose may lead to better survival for NOL-ESCC patients undergoing CCRT.