2.4. Statistical Analyses

We categorized patients into inflammation and non-inflammation groups and adopted propensity score matching (PSM) with the aim of eliminating confounding factors in the comparison of various postoperative outcomes. A logistic regression model was utilized to determine the propensity score, incorporating covariates such as age, sex, surgical approach, surgical procedure, clinical stage, and comorbidities. The nearest neighbor matching method was applied, achieving one-to-one matching with a caliper size of 0.20. Following matching, comparisons of patient characteristics and postoperative outcomes were conducted using the Mann–Whitney U-test for continuous variables and the chi-squared or Fisher’s exact test for categorical variables. Multivariate analyses, employing logistic regression, were conducted to identify risk factors for postoperative complications and calculate odds ratios (ORs). All statistical analyses were performed using EZR software Ver.1.64 (Saitama Medical Center, Jichi Medical University, Saitama, Japan), with the significance level set at p < 0.05.