【作者】 蓝永洪； 申洪； 陆药丹； 邓敏；

【Author】 LAN Yong-hong~(1,2),SHEN Hong~(1,2),LU Yao-dan~1,DENG Min~1 1.Department of Pathology,Nanfang Hospital;2.Department of Pathology,School of Basic Medical Sciences,Southern Medical University,Guangzhou,510515,China

【机构】 南方医科大学南方医院病理科； 南方医科大学基础医学院病理学系；

【摘要】 目的探讨膀胱尿路上皮癌脱落细胞的计算机判别方法,为膀胱尿路上皮细胞癌的计算机诊断奠定基础。方法取107例尿液标本中500个膀胱尿路上皮癌脱落细胞(UCC)、439个尿路上皮结构不良脱落细胞(UDC)并和386个正常尿路上皮脱落细胞(UNC),随机选取1077个细胞(占细胞80%)作为训练组,剩余的248个细胞为识别组,用计算机图像分析技术分别测试细胞的色度学和形态结构参数,对训练组细胞的测试结果进行逐步判别分析并建立判别函数,评价函数的判别效果。结果结合色度学和形态结构参数所建立判别函数的回代判别符合率为80.8%,对248个细胞(识别组)和107例标本的判别符合率分别为80.2%和92.5%,其判别效果明显优于单独基于色度学和形态结构参数所建立的判别函数(P<0.05)。结论结合色度学和形态结构参数所建立的判别函数判别效果好,为膀胱尿路上皮癌的计算机定量诊断提供了分析路径。更多还原

【Abstract】 Objective The aim of this study was to explore the method of computer discriminant on exfoliated cells in urothelium carcinoma of bladder,to establish the foundation on computer diagnosis for urothelium carcinoma of bladder.Methods The exfoliated cells in 107 urine specimens included 500 urothelium carcinoma exfoliated cells of bladder(UCC),439 urothelium dysplastic exfoliated cells(UDC)and 386 uroepithelium normal exfoliated cells(UNC).The 1077 cells(80%cells)were randomly selected as training group,and residual 248 cells as identifying group.Technique of computer image analysis was applied to test chromatics and morphological structure parameters of cells.The stepwise discriminant analysis was used to establish the discriminant function of test results in training group,and the discriminant effects of function were evaluated.Results The back discriminant coincidence rate of function combined with chromatics and morphological structure parameters was 80.8%.The coincidence rate of 248 cells(identifying group)and 107 specimens was 80.2%and 92.5%respectively.The discriminant effect was significantly better than function individually based on chromatics and morphological structure parameters(P<0.05).Conclusions The function combined with chromatics and morphological structure parameters has good discriminant effect, and it provides the analysis path of quantitative diagnosis for urothelium carcinoma of bladder.更多还原