[The impact of different comprehensive treatment models on patients with adenocarcinoma of esophagogastric junction based on propensity score matching: a single center cohort study].

To compare the prognostic influence and postoperative pathology of different comprehensive treatment models for adenocarcinoma of esophagogastric junction. Between January 2012 and December 2017, a total of 219 patients with adenocarcinoma of esophagogastric junction underwent surgery in Gastrointestinal Cancer Center, Peking University Cancer Hospital & Institute and were enrolled in this study. The clinicopathological data of these patients were collected. The patients were categorized into 3 groups according to different treatment models: surgery-first group, neoadjuvant chemotherapy (NAC) group and neoadjuvant chemoradiotherapy (nCRT) group. A trimatch propensity score analysis was applied to control potential confounders among the three groups by using R language software. A total of 7 covariates including gender, age, comorbidity, body mass index, clinical T stage, clinical N stage and Siewert type were included, and the caliper value was taken as 0.2. After matching, a total of 87 patients were included for analysis with 27 patients for each group. There were 82 males and 5 females, with a median age of 63 years (range: 38 to 76 years). The effect of preoperative treatment on postoperative tumor pathology among the three different comprehensive treatment models was explored by χ test, ANOVA or Wilcoxon rank sum test. Mann-Whitney test or χ test were used to undergo pairwise comparisons. Kaplan-Meier method and Log-rank test were used to analyze the overall survival and progression-free survival. The proportion of vascular embolism in the surgery-first group was 72.4% (21/29), which was significantly higher than NAC group (37.9% (11/29), χ=6.971, =0.008) and nCRT group (6.9% (2/29), χ=26.696, <0.01). The proportions of pathological T3-4 stage in nCRT group and NAC group were 55.2% (16/29) and 62.1% (18/29), respectively, which were significantly lower than the surgery-first group (93.1% (27/29), χ=10.881, =0.001; χ=8.031, =0.005). Compared with the NAC group (55.2% (16/29), χ=6.740, =0.009) and nCRT group (31.0% (9/29), χ=18.196, <0.01), the proportion of lymph node positivity 86.2% (25/29) were significantly higher in the surgery-first group. The 5-year overall survival rates were 62.1%, 68.6% and 41.4% for the surgery-first group, NAC group and nCRT group, respectively (χ=4.976, =0.083). The 5-year progression-free survival rates were 61.7%, 65.1% and 41.1% for the surgery-first group, NAC group and nCRT group, respectively. The differences in overall survival (χ=4.976, =0.083) and progression-free survival (χ=4.332, =0.115) among the three groups were nonsignificant. Postoperative pathology is significantly different among the three groups. Neoadjuvant chemotherapy and neoadjuvant chemoradiotherapy could decrease the proportions of vascular embolism, pathological T3-4 stage and lymph node positivity to achieve local tumor control. The prognosis of overall survival and progression-free survival are not significantly different among the three groups.