论文部分内容阅读
目的:建立应用蛋白质芯片技术检测血浆蛋白质指纹图谱的方法,探讨基于人工神经网络的血浆蛋白质指纹图谱模型在卵巢癌诊断中的应用价值。方法:应用表面增强激光解析电离飞行时间质谱仪测定卵巢癌和健康人血浆标本的蛋白质指纹图谱,并结合人工神经网络进行数据分析。将160例标本随机分成训练组120例(卵巢癌40例,健康人80例)和盲法测试组40例(卵巢癌20例,健康人20例)。利用从训练组得出的基于人工神经网络的血浆蛋白质指纹图谱模型,对40例未知血浆标本进行检测,并与CA125检测结果进行比较。结果:应用该方法对卵巢癌进行诊断的敏感性和特异性分别为95%(19/20),100%(20/20),明显高于CA125检测结果。结论:基于人工神经网络的血浆蛋白质指纹图谱模型在卵巢癌的诊断中较传统方法具有更高的敏感性和特异性,值得进一步研究与应用。
OBJECTIVE: To establish a method for the detection of plasma protein fingerprinting using protein chip technology and to explore the application value of the plasma protein fingerprinting model based on artificial neural network in the diagnosis of ovarian cancer. Methods: The protein fingerprints of plasma samples from ovarian cancer patients and healthy individuals were determined by surface enhanced laser desorption / ionization time of flight mass spectrometry. The data were analyzed by artificial neural network. 160 cases were randomly divided into training group of 120 cases (40 cases of ovarian cancer, 80 healthy subjects) and 40 cases of blinded test (20 cases of ovarian cancer and 20 healthy subjects). Using the plasma protein fingerprinting model based on artificial neural network derived from the training group, 40 cases of unknown plasma samples were tested and compared with the CA125 test results. Results: The sensitivity and specificity of this method for the diagnosis of ovarian cancer were 95% (19/20) and 100% (20/20), respectively, which were significantly higher than those of CA125. Conclusion: The model of plasma protein fingerprinting based on artificial neural network is more sensitive and specific than traditional methods in the diagnosis of ovarian cancer, which is worth further study and application.