目的: 利用蛋白质组学基质辅助激光解析离子化飞行时间质谱(matrixassisted laser desorption/ionization timeofflight mass spectrometry, MALDITOF MS)技术检测食管癌患者血清蛋白指纹图谱，建立食管癌蛋白指纹图谱诊断模型，探讨其临床应用价值。方法: 收集河北医科大学第四医院2008年5月至9月间胸外科食管癌患者血清32例和健康志愿者血清28例，采用弱阳离子交换蛋白质芯片（WCX磁珠）提纯血清蛋白，MALDITOF MS技术进行血清蛋白质质谱检测，所得结果用ZUCI蛋白芯片数据分析系统进行分析。运用遗传算法结合支持向量机运算建立食管癌蛋白指纹图谱诊断模型；将60例标本随机分成训练组和盲法测试组，训练组为21名食管癌患者和19名健康人，盲法测试组为11名食管癌患者和9名健康人，验证诊断模型的特异性和敏感性。结果:采集食管癌患者和健康对照者的血清蛋白指纹图谱，经数据对比分析找到44个有显著性差异的质荷比峰（P<0.05）；从中筛选出差异最显著的6个蛋白质荷比峰（m/z分别为2 210、2 864、6 634、4 068、2 083和8 131），并以此建立食管癌蛋白指纹图谱诊断预测模型；验证该模型诊断食管癌的特异性为88.9%、敏感度为100%。结论: 应用MALDITOF MS技术检测食管癌患者血清蛋白指纹图谱建立的食管癌蛋白指纹图谱诊断模型在食管癌的诊断中具有较高的敏感度和特异性。

Objective:To examine the serum proteomic spectra of human esophagial carcinoma by matrixassisted laser desorption/ionization timeofflight mass spectrometry (MALDITOF MS), so as to set up a diagnostic model of esophagial carcinoma and to investigate its clinical value. Methods: Thirtytwo esophagial carcinoma patients and 28 healthy controls were obtained from Fourth Affiliated Hospital of Hebei Medical University during May to September of 2008. Serum protein was extracted by weak cation exchange (WCX) protein chip system, and proteomic spectra was examined by MALDITOF MS. The obtained data were analyzed by ZUCIprotein chip data analyze system (ZUCIPCDAS) and an esophagial carcinoma diagnostic model was established by genetic arithmetic (GA) combined support vector machine (SVM). The above 60 samples were randomly divided into training set and blinding test set, with training set including 21 esophagial carcinoma patients and 19 healthy controls and blinding test set including 11 esophagial carcinoma patients and 9 healthy controls, so as to examine the specificity and sensitivity of this diagnostic model. Results: Serum proteomic spectra of esophagial carcinoma patients and healthy controls were obtained by MALDITOF MS, and m/z (mass to charge) peaks of 44 differential proteins were obtained after analyzed by ZUCIPCDAS software package (P<0.05). From which, 6 differential proteins (whose m/z peaks being 2 210, 2 864, 6 634, 4 068, 2 083 and 8 131, respectively) were selected to establish a diagnostic model of esophagial carcinoma. Specificity and sensitivity of this model in diagnosing esophagial carcinoma was 88.9% and 100%, respectively. Conclusion: An esophagial carcinoma diagnostic model has been established from serum proteomic spectra of esophagial carcinoma by MALDITOF MS and has high sensitivity and specificity to diagnose esophagial carcinoma.

科技部重大科技攻关课题 （No. 2006BAI02A07）