组织芯片技术与人工智能神经网络在大肠肿瘤诊断中的应用

目的：构建组织原位检测指标预测诊断大肠肿瘤的人工智能神经网络（ANN）模型，探讨组织芯片技术与ANN结合应用的可行性。方法：应用组织芯片技术检测ST13等8种组织原位检测指标在大肠肿瘤演进过程各阶段的表达，同时采用ANN构建相应的诊断模型。结果：采用Matlab6．5软件中提供的Kruskal-wal-lis H秩和检验函数，对这8种指标在正常大肠组织、大肠腺瘤和大肠癌组织中的阳性表达差异进行统计学检验，其中ST13、Bcl-2、Survivin和HSFlm RNA的表达，差异有统计学意义，P〈0．01；将8种指标随机组合，分别构建相应的ANN诊断模型，评价其各自的诊断效率，发现ST13、Bcl-2、Survivin与HSFl mRNA组合的ANN-BP模型预测准确率最高，其对正常大肠组织、大肠腺瘤和大肠癌训练集的预测准确率分别高达92.895％、94.163％和92.013％，对该ANN-BP网络诊断模型的盲法测试结果也分别高达85.714％、79.412％和72％。结论：组织芯片技术与ANN相结合，可以大大提高组织原位检测指标对大肠肿瘤的预测诊断效率。

Combined application of tissue microarray technique and artificial neural networks in colon tumor diagnosis

OBJECTIVE:To establish the colon tumour diagnostic models of 8 tumor related markers in tissue in situ by artificial neural network (ANN) and evaluate the feasibility of combined application of tissue microarray (TMA) technique and ANN. METHODS: Sever kinds of tumor related proteins (ST13 and so on) and HSF1 mRNA were detected by means of TMA technique, and the diagnostic models were established by ANN-BP. RESULTS: By means of Kruskal-wallis H test available in Matlab 6.5, the expression of every one of 8 tumour related markers (proteins/mRNA) was evaluated in healthy colon, colon adenoma and colon carcinoma, respectively, and the result showed that the expression in these 3 tissues of ST13, Bcl-2, Survivin and HSF1 mRNA were significantly different, P<0.01; then the random assortments of the 8 tumour relate markers were used to establish different diagnostic models, whose diagnostic sensibilities were evaluated by training and blinding test sets respectively. The diagnostic model established by the group of ST13, Bcl-2, Survivin and HSF1 mRNA was found to be the best one among all the random groups. Its training set predicted veracities were 92.895% for healthy colon tissue, 94.163% for colon adenoma, and 92.013% for colon carcinoma, meanwhile its blinding test set predicted veracities were 85.714%, 79.412% and 72%, respectively. CONCLUSION: Combined application of TMA technique and ANN can enhance the diagnostic efficiency of the tumor related markers in colon tissue in situ dramatically.