Objective To investigate the efficacy of the Ivor-Lewis cervical stapled esophagogastrostomy via thorax in the treatment of middle esophageal carcinoma.Methods The clinical data of 303 patients with middle esophageal carcinoma who were admitted to the Rugao Boai Hospital (107 patients) and the Rugao People's Hospital (196 patients) from March 2005 to March 2013 were prospectively analyzed.All the patients received Ivor-Lewis stapled cervical esophagogastrectomy (Ivor-Lewis group,151 patients) or Sweet stapled cervical esophagogastrostomy (Sweet group,152 patients) according to the admission order.The intraoperative condition,perioperative complications,lymph node dissection and postoperative follow-up of the 2 groups were analyzed.All the patients were followed up via out-patient examination till December 2012.The measurement data,enumeration data and the ranked data were analyzed using the independent samples t-test,chi-square test or Fisher exact probability and Wilcoxon rank sum test,respectively.The survival curve was drawn by the Kaplan-Meier method,and the postoperative mortality rate was analyzed using the Cox proportional hazard model.Results The operation time of the Ivor-Lewis group was (239 ± 21)minutes,which was significantly longer than (188 ± 30)minutes of the Sweet group (t =11.32,P ＜ 0.05).The surgical resection rate of the Ivor-Lewis group was 98.68％ (149/151),which was significantly higher than 92.76％ (141/152) of the Sweet group (x2 =6.45,P ＜ 0.05).The positive rate of the upper resection margin of the esophagus,postoperative morbidity rate and operative were 0.67％(1/149),10.07％ (15/149) and 0 in the Ivor-Lewis group,and 0.71％ (1/141),11.35％ (16/141) and 0.71％(1/141) in the Sweet group,with no significant difference between the 2 groups (P ＞ 0.05).The number of lymph nodes dissected from the cervical-thoracic junction and the upper abdomen were 3.6 ± 1.1 and 3.5 ± 1.1 in the Ivor-Lewis group,which were significantly greater than 2.3 ± 0.8 and 2.4 ± 0.8 in the Sweet group (Z =9.96,9.02,P ＜ 0.05).The number of positive lymph nodes dissected from the cervical-thoracic junction was 0.7 ± 1.1 in the Ivor-Lewis group,which was greater than 0.3 ± 0.6 of the Sweet group,with significant difference between the 2 groups (Z =3.26,P ＜ 0.05).Of the 290 patients who received surgical treatment,273 were followed up with the follow-up rate of 94.14％ (273/290),and the median time for follow-up was 28.0 months.The 1-,2-,3-year tumor recurrence rates were 8.21％ (11/134),19.64％ (22/112) and 29.35％ (27/92) of the Ivor-Lewis group,which was significantly lower than 19.05％ (24/126),35.24％ (37/105) and 44.19％(38/86) of the Sweet group (x2=6.55,7.33,5.03,P ＜ 0.05).There were significant differences in the 1-,2-,3-year locoregional recurrence rate of the lymph nodes between the 2 groups (x2 =7.03,9.68,6.87,P ＜0.05).The 1-,2-,3-year accumulative survival rates of the Ivor-Lewis group were 90.30％ (121/134),80.36％(90/112) and 71.74％ (66/92),which were significantly higher than 80.95％ (102/126),59.05％ (62/105)and 51.16％ (44/86) of the Sweet group (x2=4.65,1 1.73,7.97,P ＜ 0.05).Conclusion Ivor-Lewis stapled cervical esophagogastrostomy via thorax has advantages of high resection rate,better safety and better quality of life of patients,and it could be an optimized design of the treatment for patients with middle esophageal cancer without intumescent lymph node of neck.