Statistical analysis

Continuous data were compared using the Mann-Whitney U test. The categorical variables were compared using the chi-square test or Fisher’s exact test. Statistical significance was set P<0.05. The overall survival, RFS, and CSS were assessed by plotting Kaplan-Meier curves. Cox’s proportional hazard model was used to estimate the hazard ratios (HRs) with 95% confidence intervals (CIs).

Propensity score matching was conducted to reduce confounding effects. The propensity score is the probability that an individual patient would have been assigned to undergo D2 dissection, conditional on observed covariates. Propensity scores were estimated using a logistic regression model based on age, sex, body mass index, American Society of Anesthesiologists Physical Status (ASA-PS), clinical T, clinical N, and procedures. D2 cases were matched 1:1 to Non-D2 cases with similar pre-interventional probability without a replacement using a 0.2 caliper width. Although there is no consensus on which thresholds for standardized differences should be used to detect the residual imbalance across groups in matched samples, an absolute standardized difference of more than 0.25 was considered as a sign of imbalance according to the some reports (20-22). The resulting score-matched pairs were analyzed subsequently. All statistical analyses were performed using the JMP 14 software (SAS institute, Cary, NC, USA).