摘要
目的探讨早期胃癌内镜黏膜下剥离术(ESD)术后迟发性出血的评分预测模型(BEST-J评分)验证及新模型开发。方法回顾性分析2017-01至2022-12联勤保障部队第910医院消化内科接受ESD治疗的早期胃癌患者临床资料。根据是否发生出血将全部患者分为迟发性出血组和无出血组。对两组患者临床资料进行单因素比较和多因素logistic回归分析,确定影响迟发性出血的因素,并构建新的迟发性出血预测模型。采用受试者工作特征(ROC)曲线评价新预测模型与BEST-J评分的预测效能。结果共纳入428例接受ESD的早期胃癌患者,其中迟发性出血26例(6.07%)。与无出血组相比,迟发性出血组患者的平均BEST-J评分更高(P<0.001),高危和极高危患者比例更高(P=0.013)。多因素logistic回归分析显示,吸烟史(OR=2.219,95%CI:1.068~4.610,P=0.040)、肿瘤最大直径>20 mm(OR=2.563,95%CI:1.244~5.283,P=0.010)、切除多发性肿瘤(OR=3.179,95%CI:1.134~8.913,P=0.033)和APTT延长(OR=5.959,95%CI:1.867~19.015,P=0.002)是迟发性出血的独立危险因素,而术中使用止血夹(OR=0.312,95%CI:0.118~0.826,P=0.021)是保护因素。构建迟发性出血风险预测模型为:Logit(P)=-3.867+0.797×(吸烟史)-1.165×(术中使用止血夹)+1.157×(切除多个肿瘤)+0.941×(肿瘤>20 mm)+1.785×(APTT延长)。ROC曲线分析表明,基于影响因素构建的新预测模型曲线下面积(AUC)优于BEST-J评分(0.622 vs.0.809,P<0.001)。结论基于吸烟史、肿瘤大小>20 mm、术中使用止血夹、切除多发肿瘤和APTT延长因素的新预测模型较BEST-J评分更适合中国早期胃癌患者ESD术后出血风险预测。
Objective To evaluate the prediction model of delayed bleeding after endoscopic submucosal dissection(ESD)for early gastric cancer and to develop a new model.Methods A retrospective analysis was conducted on clinical data of patients with early gastric cancer who received ESD treatment in Gastroenterology Department of the 910th Hospital of PLA Joint Logistics Support Force from January 2017 to December 2022.All patients were divided into delayed bleeding group and non-bleeding group according to whether bleeding occurrence.Univariate and multivariate logistic regression analysis were conducted on clinical data of the two groups to identify factors affecting delayed bleeding,and a new delayed bleeding prediction model was constructed.The predictive performance of the new model and the BEST-J score was evaluated using the receiver operating characteristic(ROC)curve.Results A total of 428 patients with early gastric cancer who received ESD were included,with 26 cases(6.07%)of delayed bleeding.Compared with the non-bleeding group,the delayed bleeding group had higher mean BEST-J score(P<0.001)and higher proportion of high-risk and very high-risk patients(P=0.013).Multivariate logistic regression analysis showed that smoking history(OR=2.219,95%CI:1.068-4.610,P=0.040),tumor maximum diameter>20 mm(OR=2.563,95%CI:1.244-5.283,P=0.010),resection of multiple tumors(OR=3.179,95%CI:1.134-8.913,P=0.033),and prolonged APTT(OR=5.959,95%CI:1.867-19.015,P=0.002)were independent risk factors,while intraoperative use of hemostatic clips(OR=0.312,95%CI:0.118-0.826,P=0.021)was a protective factor.The risk prediction model for delayed bleeding was constructed as follows:Logit(P)=-3.867+0.797×(smoking history)-1.165×(intraoperative use of hemostatic clips)+1.157×(resection of multiple tumors)+0.941×(tumor>20 mm)+1.785×(APTT prolongation).ROC curve analysis showed that the new prediction model based on influencing factors had a larger area under the curve(AUC)than the BEST-J score(0.622 vs 0.809,P<0.001).Conclusions The new predic
作者
吴雅鋆
曾清芳
连丽凤
施亚珠
黄永德
WU Yayun;ZENG Qingfang;LIAN Lifeng;SHI Yazhu;HUANG Yongde(Department of Gastroenterology,the 910th Hospital of PLA Joint Logistics Support Force,Quanzhou 362000,China)
出处
《武警医学》
CAS
2024年第7期572-577,共6页
Medical Journal of the Chinese People's Armed Police Force
基金
福建省自然科学基金项目(2023J011831)
泉州市科技计划项目(2018N33S)
关键词
内镜黏膜下剥离术
早期胃癌
迟发性出血
预测
endoscopic submucosal dissection
early gastric cancer
delayed bleeding
forecast
