Abstract: Objective To investigate the clinical value of computer tomography (CT)-based three-section formula in identification of Siewert types of adenocarcinoma of esophagogastric junction (AEG). Methods The retrospective and descriptive study was conducted. The clinicopathological data of 62 AEG patients who were admitted to two medical centers, including 33 patients from Peking University Cancer Hospital and 29 patients from the First Affiliated Hospital of Amy Medical University, between January 2019 and January 2021 were collected. There were 53 males and 9 females, aged (66±9)years. All patients underwent CT examination to obtain the coronal and axial images and determine the upper and lower edges of the tumor and the esophagogastric junction, which were imported into the formula for Siewert classification. Observation indicators: (1) results of CT examination and pathological examination; (2) consistence of judgements for CT examination between doctors; (3) consistence of judgements between CT examination and pathological examination. Results of pathological examination came from intraoperative surgical observation and postoperative histopathological examination. Measurement data with normal distribution were represented as Mean±SD, and count data were described as absolute numbers or percentages. The consistency coefficient Kappa (κ) was used to evaluate the consistency of diagnosis between resear-chers. The κ≥0.75 was regarded as excellent consistency, 0.40<κ<0.75 as good consistency, κ ≤0.40 as poor consistency. Wilcoxon rank sum test was used to evaluate the statistical difference between results of the CT-based three-section formula method and the pathological examination. Taking the results of histopathological examination as standard, the sensitivity, specifi-city, accuracy and 95% confidence interval of the CT-based three-section formula method were calculated. Results (1) Results of CT examination and pathological examination: all the 62 patients underwent CT examination successfully to obtain the coronal and axial images and determine the upper and lower edges of the tumor and the esophagogastric junction, which were used for Siewert classification. There were 3 cases with AEG of Siewert type Ⅰ, 47 cases with AEG of Siewert type Ⅱ and 12 cases with AEG of Siewert type Ⅲ according to doctor's judgement from the Peking University Cancer Hospital, and there were 3 cases with AEG of Siewert type Ⅰ, 49 cases with AEG of Siewert type Ⅱ and 10 cases with AEG of Siewert type Ⅲ according to doctor's judgement from the First Affiliated Hospital of Amy Medical University, respectively. After arbitration, there were 2 cases with AEG of Siewert type Ⅰ, 48 cases with AEG of Siewert type Ⅱ and 12 cases with AEG of Siewert type Ⅲ determined by the CT based three-section formula. There were 7 cases of stage T1, 10 cases of stage T2, 24 cases of stage T3, 14 of stage T4a and 7 cases of stage T4b of pathological T staging. There were 2 cases with AEG of Siewert type Ⅰ, 48 cases with AEG of Siewert type Ⅱ and 12 cases with AEG of Siewert type Ⅲ determined by pathological examination. (2) Consistence of judgements for CT examination between doctors: the consistency of Siewert classification determined by CT-based three-section formula between 2 doctors was good (κ=0.74, P<0.001). (3) Consistence of judgements between pathological examination and CT examination: with Siewert classification determined by pathological examination as reference, the accuracy of Siewert classification determined by CT based three-section formula was 90.3%(κ =0.73, P<0.001). The sensitivity and specificity of CT-based three-section formula were 66.7%(95% confidence interval as 20.8%?93.9%) and 100.0%(95% confidence interval as 93.9% ? 100.0%) for AEG of Siewert type Ⅰ, 97.7%(95% confidence interval as 88.2%?99.6%) and 72.2%(95% confidence interval as 49.1%?87.5%) for AEG of Siewert type Ⅱ, 73.3%(95% confidence interval as 48.0% ? 89.1%) and 97.9%(95% confidence interval as 88.9% ? 99.9%) for AEG of Siewert type Ⅲ, respectively. Conclusion The CT-based three-section formula can be used for identification of Siewert types of AEG, with a high accuracy.