肺癌血清蛋白质指纹图谱诊断模型的建立

目的建立肺癌血清蛋白质指纹图谱诊断模型,探讨其对肺癌的诊断价值。方法收集127例肺癌患者、108例正常人血清标本,随机分成训练组(肺癌89例、正常人68例)与测试组(肺癌38例、正常人40例),应用表面增强激光解吸/电离-飞行时间-质谱(SELDI-TOF-MS)检测其蛋白质指纹图谱。用BioMarker Pattern软件分析训练组中肺癌与正常人血清中的差异蛋白后,建立肺癌诊断的最优分类树模型,然后用测试组血清标本盲法验证该模型的诊断效率。结果分类树诊断模型由质荷比分别为5 808、5 971和7 780的3种蛋白质组成。盲法验证显示,该模型诊断肺癌的准确率为92.3%,敏感性与特异性分别为92.1%(35/38)和92.5%(37/40)。结论肺癌血清蛋白质指纹图谱诊断模型较以往的传统方法具有更高的敏感性与特异性,值得进一步研究。

Establidmlent of diagnostic model for lung cancer by protein fingerprinting

Objective To establish the diagnostic model for lung cancer by serum protein fingerprinting and evaluate its efficacy on the diagnosis of lung cancer.Methods Serum samples from 127 patients of lung cancer and 108 healthy people were randomly divided into training group(89 cases of lung cancer and 68 healthy people) and test group(38 cases of lung cancer and 40 healthy people).Their sera protein fingerprinting were read by surface-enhanced laser desorption/ionization-time of flight-mass spectrometry(SELDI-TOF-MS).Using BioMarker Pattern software to analyze the distinct proteins between lung cancer and healthy people of training set and create a best decision classification tree model for diagnosing lung cancer.Then the model was blindly validated by the protein fingerprinting of the test group.Results The best decision classification tree model was successfully established,which including three proteins with the mass-to-charge ratio of 5 808,5 971 and 7 780.Blinded validation suggested that the accuracy for diagnosing lung cancer was 92.3%,and the sensitivity and specificity were 92.1%(35/38) and 92.5%(37/40),respectively.Conclusion The serum protein fingerprinting diagnostic model for lung cancer yields rather higher sensitivity and specificity than those of other traditional methods,which deserves further research.