食管癌血清差异蛋白质组测定结果分析

Analysis of detection results of different serum proteomes in esophageal cancer

目的用表面增强激光解析电离飞行时间质谱技术(SELDI-TOF-MS)寻找食管癌(EC)患者血清中差异蛋白质组,并建立诊断模型,为其广泛用于临床诊断积累数据资料。方法用SELDI芯片检测食管癌及其相关人群的血清,Biomarker Wiz-ard Software软件筛选出差异蛋白,选择具有显著差异的蛋白质组建立人工神经网络诊断模型,经盲法验证后使用SPSS软件分析各个差异蛋白以及联合人工神经网络技术后的诊断效率,并鉴定相关蛋白。结果筛选出显著差异(P<0.01)的蛋白质组如下:在食管癌患者血清中表达增高蛋白组[(5 017.6±4.89)、(7 458.5±6.49)、(7 908.1±7.80)、(8 111±8.45)、(8 577±7.80)kD],表达降低蛋白组[(7 748.3±9.10)、(5 890.9±7.32)、(4 213.8±5.93)kD]。差异蛋白质组建立食管癌人工神经网络筛查模型和诊断模型,其灵敏度和特异性超过90%,盲法验证的效果好;人工神经网络预测输出值构成的ROC曲线下面积分别为81.6%和82.1%,经查阅数据库鉴定发现两种蛋白质。结论血清蛋白质谱指纹图结合人工神经网络技术进行蛋白质组学数据挖掘所建模型诊断价值较大,与临床诊断符合程度良好。

Objective To find differential serum protomes in esophageal cancer by enhanced laser desorption ionization time of flight mass spectrometry（SELDI-TOF-MS） to establish diagnostic models for accumulating the data information for wide use in clinical diagnostic.Methods To detect the serum of the patients with esophageal cancer and related groups by SELDI chip,filter out different proteomes with the Biomarker Wizard Software.Then to develop the artificial neural network diagnostic models using significant different proteomes and use SPSS software to analyze diagnostic efficiency of significant difference in proteomes and the combined artificial neural network after verified by the blind.Finally,to identify related proteins.Results The selected different proteomes（P0.01） were as follows.In esophageal cancer,the intensity increasing proteomes were 5 017.6±4.89,7 458.5±6.49,7 908.1±7.80,8 111±8.45,8 577±7.80 kD,lowerly expressed proteomes were 7 748.3±9.10,5 890.9±7.32,4 213.8±5.93 kD.Built the screening and diagnosis models of artificial neural network by different proteomes of esophageal cancer.Sensitivity and specificity were over 90%.After blinding test,they were effective.The area under the ROC curve consisting by the output value of artificial neural network were 81.6% and 82.1%,searching the related databases,two proteins were found.Conclusion The value of dioglositic models are useful by associating serum protein fingerprint spectrums with artificial neural network technology in proteomics data mining and are accord with the clinical diagnosis.