AUTHOR=Xu Yuting , Luo Lijie , Feng Xingyu , Zheng Yensheng , Chen Tao , Zhou Rui , Li Yong , Li Guoxin , Wang Wei , Xiong Wenjun TITLE=Nomogram for Predicting Risk of Esophagogastric Junction (EGJ) Resection During Laparoscopic Resection of Gastrointestinal Stromal Tumors in EGJ: A Retrospective Multicenter Study JOURNAL=Frontiers in Surgery VOLUME=8 YEAR=2021 URL=https://www.frontiersin.org/journals/surgery/articles/10.3389/fsurg.2021.712984 DOI=10.3389/fsurg.2021.712984 ISSN=2296-875X ABSTRACT=

Background: The established criteria for determining whether to excise the cardia during laparoscopic surgery for gastrointestinal stromal tumors in the esophagogastric junction (EGJ-GISTs) remain controversial. This retrospective multicenter study was conducted to develop a nomogram for predicting the risk of the cardia excision during laparoscopic surgery for EGJ-GISTs.

Material and Methods: We reviewed data from 2,127 gastric-GISTs (g-GISTs) patients without distant metastases in four hospital between June 2012 and June 2020. Of those, according to the including criteria, 184 patients [Guangdong Provincial Hospital of Chinese Medicine (n = 81), Nanfang Hospital of Southern Medical University (n = 60), Guangdong General Hospital (n = 34), and The Third Affiliated Hospital of Southern Medical University (n = 9)] with EGJ-GISTs were identified and included in this study. Factors contributing to risk of cardia excision were identified and used to create a nomogram. Nomogram performance was assessed using a bootstrapped concordance index (c-index) and calibration plots.

Results: According to the multivariate analysis, the distance from the margin of the tumor to the esophagogastric line (EG-line) (cm) (OR = 0.001, 95% CI: 0.00001~0.056, P = 0.001) and tumor size (cm) (OR = 14.969, 95% CI: 1.876~119.410, P = 0.011) were significantly related to likelihood of cardia structure excision in laparoscopic surgery for EGJ-GISTs. These two factors were used to generate a nomogram for predicting risk of cardia excision using a logistic regression model; a bootstrapped C-index of 0.988 (calibrated C-index = 0.987) indicated strong predictive ability, with broad calibration.

Conclusions: This nomogram based on distance from tumor margin to EG-line and tumor size may serve as a tool for predicting risk of cardia damage during laparoscopic removal of EGJ-GISTs to aid in selection of surgical methods and preoperative neoadjuvant therapy.