摘要
目的:探讨胸水中腺癌细胞(ACC)和反应性间皮细胞(RMC)的计算机判别方法 ,为胸水中ACC计算机诊断奠定基础。方法:取胸水涂片中764个ACC(44例)和709个RMC(39例),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,随机选取约80%细胞作为训练组,剩余的20%细胞为测试组,对训练组的测试结果进行逐步判别分析并建立判别函数,用测试组评价函数的判别效果。结果:逐步判别分析显示,RP、BP、Rn、Bn、dmaj,c、dmin,c、An、Dn、Cn进入判别函数,符合率为84.5%,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,为胸水中腺癌细胞的计算机定量诊断提供了分析路径。
Objective:To explore the method of computer discriminant on adenocarcinoma cells(ACC) and reactive mesothelial cells(RMC) amd establish the foundation on computer diagnosis for ACC in hydrothorax.Methods:Totally764 ACC of 44 cases and 709 RMC of 39 cases were collected from hydrothorax smears.Technique of computer image analysis was applied to test chromatics and morphological structure parameters of above cells.Approximately 80% cells were randomly selected as train group,and residual 20% cells as test group,the stepwise discriminant analysis was used in train group to establish the discriminant function,and the discriminant effects of function in test group were evaluated.Results:Stepwise discriminant showed that RP,BP,Rn,Bn,dmaj,c、dmin,c、An、 Dn、Cn entered discriminant function,and its coincidence rate was 84.5%.The discriminant effect was significantly better than that in functions individually based on chromatics and morphological structure parameters(P 0.05).Conclusion:The function combined with chromatics and morphological structure parameters has good discriminant effect,and it provides the analysis path of computer quantitative diagnosis for ACC in hydrothorax.
出处
《南京医科大学学报(自然科学版)》
CAS
CSCD
北大核心
2013年第8期1152-1154,共3页
Journal of Nanjing Medical University(Natural Sciences)
基金
海南省自然科学基金资助(812194)
关键词
胸水
腺癌细胞
反应性间皮细胞
色度学
形态结构
计算机判别
hydrothorax
adenocarcinoma cell
reactive mesothelial cell
chromatics
morphological structure
computer discriminant