胃粘膜异型增生多项生物标志定量观察及癌变概率模糊判别研究

选择具有完整随访结果的６２例胃粘膜异型增生，其中癌变组及未癌变组各３１例随访前、后的标本，采用ＤＮＡ、核仁组成区嗜银蛋白（ＡｇＮＯＲ）、胃癌相关抗原（ＭＧ７Ａｇ）等生物学标志进行定量图象分析，并运用模糊数学的理论和判别方法，探索各单项指标及其多项指标模糊判别对胃粘膜异型增生癌变概率预测的意义，建立判别模式及计算机软件。结果：ＤＮＡ含量测定的均值、ＡｇＮＯＲ测定的均数Ｘ均径、ＭＧ７Ａｇ测定的灰度值等三个参数可分性测度值最大，用此三个最佳参数组合，建立了胃粘膜异型增生癌变概率模糊判别模式及计算软件系统。以癌变概率＞０。５作为癌变的判别标准，进行样本的回代验证，结果胃粘膜异型增生随访癌变组与未癌变组预测判别符合率分别为９３．５％和９０．３％。表明建立的判别模式具有预测价值，可供临床应用。

Guantitative study of multiple biomarkers and fuzzy discrimination in patientswith gastric of malignacy dysplasia

Sixty-two cases of gastric dysplasia with complete follow-up datawere analyzed,multiple biomarkers(DNA,AgNOR,MG7Ag)were used as parameters inquantitative computed-processed imaging analysis. Those were divided into 2 groups,i.e. gastric dysplasia with cancer and without,each constituted 31 cases. Theories offuzzy mathematics and method of pattern discrimination were employed to evaluate thesignificance of single and multiple indexes in prediction of probability of cancer and indevelopment of discriminating computer software.Results: The mean value of DNAcontent,the value of mean number times mean diameter of AgNOR and the grey valueof MG7Ag were three most distinctive valuable indices. With this grouping,we set upthe probability of cancer and fuzzy discrimination model of gastric dysplasia and thecomputer software. Value>0.5 was taken as the discriminating criterion of probability0f cancer,the same samples were rechecked by the above-mentioned model,the concor-dance rates of cancer group and non-cancer group were 93.5%and 90.3%,respectively.These indicate this discrimination model is of predictive and practical value.