基于卷积神经网络的白光内镜下早期胃癌检测  

Detection of early gastric cancer in white light imagings based on region-based convolutional neural networks

作　　者：晋晶 张倩倩[1,2] Bill Dong 马涛 王曦[1,2] 梅雪灿 宋绍方 彭杰 吴艾久 董兰芳 孔德润[1,2] Jin Jing;Zhang Qianqian;Bill Dong;Ma Tao;Wang Xi;Mei Xuecan;Song Shaofang;Peng Jie;Wu Aijiu;Dong Lanfang;Kong Derun(Dept of Gastroenterology,The First Affiliated Hospital of Anhui Medical University,Hefei 230022;Key Laboratory of Digestive Diseases of Anhui Province,Hefei 230022;School of Computer Science and Technology,High-tech Campus,University of Science and Technology of China,Hefei 230027;Hefei Zhongna Medical Instrument Co.,Ltd,Hefei 230088)

机构地区：[1]安徽医科大学第一附属医院消化内科,合肥230022 [2]安徽省消化病重点实验室,合肥230022 [3]中国科学技术大学计算机科学与技术学院,合肥230027 [4]合肥中纳医学仪器有限公司,合肥230088

出　　处：《安徽医科大学学报》2023年第2期285-291,共7页Acta Universitatis Medicinalis Anhui

基　　金：安徽省重点研究与开发计划项目(编号:2022e07020048)。

摘　　要：目的开发一种基于区域卷积神经网络(Mask R-CNN)的内镜下自动检测早期胃癌(EGC)系统。方法首先从安徽医科大学第一附属医院获得3579张和892张EGC白光图像(WLI),分别进行训练和测试。随后前瞻获取10个WLI实时视频以测试动态性能。另外再随机选取400张WLI图像,用于Mask R-CNN系统和内镜医师对照。诊断能力以准确率、敏感度、特异度、阳性预测值(PPV)和阴性预测值(NPV)进行评估。结果在WLI图像诊断中,Mask R-CNN系统准确率、敏感度和特异度分别为90.25%、91.06%和89.01%,与病理诊断差异无统计学意义。在WLI视频中,诊断EGC的准确率为90.27%,实时测速可达35帧/s。在对照实验中,Mask R-CNN系统的敏感度明显高于高年资组医师(93.00%vs 80.20%,χ^(2)=7.059,P<0.001),特异度高于低年资组医师(82.67%vs 71.87%,χ^(2)=9.955,P<0.001),总体准确率高于中年资组医师(85.25%vs 78.00%,χ^(2)=7.009,P<0.001)。结论在WLI下,Mask R-CNN系统检测EGC的性能良好,在实际临床应用中有较大潜力。Objective To develop an endoscopic automatic detection system in early gastric cancer(EGC)based on a region-based convolutional neural network(Mask R-CNN).Methods A total of 3579 and 892 white light images(WLI)of EGC were obtained from the First Affiliated Hospital of Anhui Medical University for training and testing,respectively.Then,10 WLI videos were obtained prospectively to test dynamic performance of the RCNN system.In addition,400 WLI images were randomly selected for comparison with the Mask R-CNN system and endoscopists.Diagnostic ability was assessed by accuracy,sensitivity,specificity,positive predictive value(PPV),and negative predictive value(NPV).Results The accuracy,sensitivity and specificity of the Mask R-CNN system in diagnosing EGC in WLI images were 90.25%,91.06% and 89.01%,respectively,and there was no significant statistical difference with the results of pathological diagnosis.Among WLI real-time videos,the diagnostic accuracy was 90.27%.The speed of test videos was up to 35 frames/s in real time.In the controlled experiment,the sensitivity of Maks R-CNN system was higher than that of the experts(93.00%vs 80.20%,χ^(2)=7.059,P<0.001),and the specificity was higher than that of the juniors(82.67%vs 71.87%,χ^(2)=9.955,P<0.001),and the overall accuracy rate was higher than that of the seniors(85.25%vs 78.00%,χ^(2)=7.009,P<0.001).Conclusion The Mask R-CNN system has excellent performance for detection of EGC under WLI,which has great potential for practical clinical application.

关 键 词：人工智能 区域卷积神经网络 白光内镜 早期胃癌 

分 类 号：R573.9[医药卫生—消化系统]