人工智能辅助消化内镜用于胃癌及癌前病变鉴别诊断临床价值分析  

Clinical value of artificial intelligence assisted digestive endoscopy in the differential diagnosis of gastric cancer and precancerous lesions

作　　者：王君[1] 晁帅恒 WANG Jun;CHAO Shuai-heng

机构地区：[1]河南科技大学附属许昌市中心医院消化内镜中心,河南许昌461000

出　　处：《中国疗养医学》2023年第6期656-659,共4页Chinese Journal of Convalescent Medicine

基　　金：2020年度河南省医学科技攻关计划(LHGJ20200982),人工智能肠镜辅助诊断技术提升肠道腺瘤检出率。

摘　　要：目的探究人工智能辅助消化内镜用于胃癌及癌前病变鉴别诊断的临床价值。方法回顾河南科技大学附属许昌市中心医院2020年1月至2021年12月收治的160例疑似胃癌及癌前病变患者资料,所有患者均接受常规内镜检查、人工智能辅助消化内镜检查、活检病理学诊断,分析常规内镜检查、人工智能辅助消化内镜检查对胃黏膜病变检出效能及胃癌与癌前病变鉴别诊断效能。结果活检病理学诊断160例患者,胃黏膜病变阳性140例,包括38例胃癌,102例胃上皮内瘤变(GIN);常规消化内镜检出胃黏膜病变123例,包括30例胃癌,93例GIN,其诊断胃黏膜病变的敏感度、特异度、准确率、阳性预测值、阴性预测值分别为83.57%、70.00%、81.88%、95.12%、37.84%,其鉴别胃癌与GIN的敏感度、特异度、准确率、阳性预测值、阴性预测值分别为71.05%、96.47%、88.62%、90.00%、88.17%;人工智能辅助消化内镜检出胃黏膜病变132例,包括35例胃癌,97例GIN,其诊断胃黏膜病变的敏感度、特异度、准确率、阳性预测值、阴性预测值分别为92.86%、90.00%、92.50%、98.48%、64.29%,其鉴别胃癌与GIN的敏感度、特异度、准确率、阳性预测值、阴性预测值分别为86.84%、97.87%、94.70%、94.29%、94.85%;人工智能辅助消化内镜诊断胃黏膜病变的敏感度、准确率、阴性预测值高于常规消化内镜,差异有统计学意义(P<0.05),两种检查方法鉴别诊断胃癌及GIN的效能参数比较差异无统计学意义(P>0.05)。结论人工智能辅助消化内镜有助于提高胃黏膜病变检出效能,有效辅助胃癌及GIN鉴别诊断。Objective To explore the clinical value of artificial intelligence assisted digestive endoscopy in the differential diagnosis of gastric cancer and precancerous lesions.Methods The data of 160 patients with suspected gastric cancer and precancerous lesions admitted to Xuchang Central Hospital Affiliated to Henan University of Science and Technology from January 2020 to December 2021 were reviewed.All patients underwent conventional endoscopy,artificial intelligence assisted digestive endoscopy,and biopsy pathology diagnosis.The detection efficiency of conventional endoscopy and artificial intelligence assisted digestive endoscopy for gastric mucosal lesions and the differential diagnosis efficiency of gastric cancer and precancerous lesions were analyzed.Results Among 160 patients diagnosed by biopsy pathology,140 were positive for gastric mucosal lesions,including 38 cases of gastric cancer and 102 cases of gastric intraepithelial neoplasia(GIN);Conventional digestive endoscopy detected 123 cases of gastric mucosal lesions,including 30 cases of gastric cancer and 93 cases of GIN.The sensitivity,specificity,accuracy,positive predictive value,and negative predictive value for diagnosing gastric mucosal lesions were 83.57%,70.00%,81.88%,95.12%,and 37.84%,respectively.The sensitivity,specificity,accuracy,positive predictive value,and negative predictive value for distinguishing gastric cancer from GIN were 71.05%,96.47%,88.62%,90.00%,and 88.17%,respectively;Artificial intelligence assisted digestive endoscopy detected 132 cases of gastric mucosal lesions,including 35 cases of gastric cancer and 97 cases of GIN.The sensitivity,specificity,accuracy,positive predictive value,and negative predictive value for diagnosing gastric mucosal lesions were 92.86%,90.00%,92.50%,98.48%,and 64.29%,respectively.The sensitivity,specificity,accuracy,positive predictive value,and negative predictive value for distinguishing gastric cancer from GIN were 86.84%,97.87%,94.70%,94.29%,and 94.85%,respectively;The sensitivity,accuracy,and ne

关 键 词：人工智能辅助消化内镜 胃癌 癌前病变 鉴别诊断 

分 类 号：R735.2[医药卫生—肿瘤]