期刊文献+

胃ESD术中出血的危险因素分析及Logistic回归模型建立 被引量:4

Logistic regression Analysis of risk factor of bleeding during gastric ESD operation and establishment of Logistic regression model
下载PDF
导出
摘要 目的 分析胃内镜黏膜下剥离术(ESD)术中出血的危险因素,并建立Logistic回归模型。方法 选择接受胃ESD术治疗患者96例。采用Logistic回归分析其独立危险因素并构建Logistic回归模型。结果 单因素、多因素Logisitic回归分析显示,病理类型早期胃癌、病变直径>5 cm、病变部位贲门-胃底、术中粘连是胃ESD术中出血的独立危险因素(P<0.05)。回归模型拟合优度良好,预测胃ESD术中出血的AUC为0.892,灵敏度、特异度分别为90.00%、77.27%。结论 病理类型为早期胃癌、病变直径>5 cm、病变部位贲门-胃底、术中粘连是胃ESD术中出血的独立危险因素。回归模型拟合优度良好,对危险因素加以防控可以降低术中出血发生率。 Aim To analyze the risk factors of bleeding during gastric endoscopic submucosal dissection(ESD), and construct a Logistic regression model. Methods 96 patients who received gastric ESD were selected. Independent risk factors were analyzed by Logistic regression, and Logistic regression model was established. ResultsUnivariate analysis and multivariate Logisitic regression analysis showed that early cancer, lesion diameter >5 cm, lesions located in cardia-fundus and intraoperative adhesion were independent risk factors of intraoperative bleeding in gastric ESD(P<0.05). The regression model has good goodness of fit, and the AUC of predicting gastric ESD intraoperative bleeding is 0.892, sensitivity and specificity were 90.00%, 77.27%. Conclusion Early cancer, lesion diameter >5 cm, lesions located in cardia-fundus and intraoperative adhesion are independent risk factors of intraoperative bleeding in gastric ESD. The regression model has good goodness of fit, and prevention and control of risk factors can reduce the incidence of intraoperative bleeding.
作者 张炜 ZHANG Wei(Ningde Municipal of Ningde Normal University,Ningde 352100,Fujian,China)
出处 《中南医学科学杂志》 CAS 2023年第1期77-79,106,共4页 Medical Science Journal of Central South China
基金 福建省自然科学基金面上项目(2021JOL1161)。
关键词 内镜黏膜下剥离术 早期胃癌 癌前病变 危险因素 endoscopic submucosal dissection early gastric cancer precancerous lesion risk factor
  • 相关文献

参考文献6

二级参考文献31

共引文献269

同被引文献26

引证文献4

二级引证文献4

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部