SELDI-TOF-MS技术在结直肠腺瘤血清蛋白质指纹图谱筛查模型建立中的应用

Application of SELDI-TOF-MS technology on serum protein fingerprinting screening in colorectal adenoma

目的应用表面增强激光解吸/离子化飞行时间质谱(SELDI-TOF-MS)技术建立结直肠腺瘤(CRA)患者血清蛋白质指纹图谱筛查模型。方法随机选取CRA 42例、结直肠良性疾病(CBD)36例及正常人(HC)44例的血清标本组成建模组,应用SELDI-TOF-MS检测其蛋白质指纹谱,采用Biomarker Wizard及Biomarker Patterns软件分析建模组中各类人群血清中的差异蛋白后,建立CRA筛查最优分类树模型;另随机抽取血清标本70例(CRA及CBD各20例、HC 30例)组成测试组,盲法验证该模型对CRA的筛查效能。结果成功建立了结直肠腺瘤筛查分类树模型。测试模式下,该模型的诊断准确率87.70%、灵敏度71.43%、总特异度96.25%、阳性预测值90.91%。盲法验证该模型诊断准确率88.57%,灵敏度60.0%,总特异度100.0%,阳性预测值100.0%。结论应用SELDI-TOF-MS技术成功建立了CRA筛查模型,该模型敏感性与特异性较高。

Objective To establish serum protein fingerprinting screening model for colorectal adenoma（CRA） with surface-enhanced laser desorption/ionization-time of flight-mass Spectrometry（SELDI-TOF-MS） technology.Methods 122 serum samples including 42 cases of CRA,36 cases of colorectal benign diseases（CBD） and 44 cases of healthy control（HC）,were randomly collected and formed the training group.Their serum protein fingerprintings were read by SELDI-TOF-MS.Biomarker Wizard and Biomarker Patterns software were separately utilized to analyze the distinct proteins between CRA,CBD and HC,and to create the best decision classification tree model for colorectal adenoma screening.Then the model was blindly validated by the protein fingerprinting of the test group（CRA 20,CBD 20 and HC 30） which was also randomly selected at the same period.Results The best decision classification tree model for CRA screening was successfully established.In the testing mode,the accuracy of prediction of CRA was 87.70%,the positive predictive value was 90.91%,the sensitivity and specificity were 71.43% and 96.25%,respectively.The values of blinded validation were 88.57%,100%,60% and 100%,respectively.Conclusion The serum protein fingerprinting screening model for CRA yields a high sensitivity and specificity,which deserves further study.