基于人工神经网络的血清蛋白质指纹图谱模型在肝癌诊断中的应用  

Using ANN and serum protein pattern models in liver cancer diagnosis

作　　者：王家祥[1] 张波[2] 余捷凯[1] 刘建[1] 杨美琴[1] 郑树[1] 

机构地区：[1]浙江大学二院肿瘤研究所,杭州310000 [2]郑州大学第一附属医院外科

出　　处：《中华医学杂志》2005年第3期189-192,共4页National Medical Journal of China

摘　　要：目的建立蛋白质芯片技术检测血清蛋白质指纹图谱的方法,探讨基于人工神经网络的血清蛋白质指纹图谱模型在肝癌诊断中的应用价值。方法应用蛋白质指纹图谱分析仪(表面增强激光解析电离飞行时间质谱仪,SELDITOFMS),测定106例肝癌、肝硬化患者和健康人血清标本的蛋白质指纹图谱并结合人工神经网络方法进行数据的分析。将106例标本随机分成训练组70例(肝癌35例,肝硬化14例,健康人21例)和盲法测试组36例(肝癌17例,肝硬化8例,健康人11例)。利用从训练组得出的基于人工神经网络的血清蛋白质指纹图谱模型,对36例未知血清进行检测,并与甲胎蛋白(AFP)检测结果进行比较。结果应用该方法对肝癌进行诊断的准确率、敏感性和特异性分别为917%(33/36)、882%(15/17)和946%(18/19),明显高于AFP检测结果。结论基于人工神经网络的血清蛋白质指纹图谱模型在肝癌的诊断中较以往的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 set up a method for the detection of the serum protein fingerprint pattern by using the protein chip technology for exploration of serum protein fingerprint pattern models based on the artificial neural network in diagnosis of liver cancer.Methods One hundred and six serum samples form subjects with liver cancer, hepatocirrhosis, and healthy individuals were detected with protein biochip surface enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) for protein fingerprint pattern, and analyzed with the artificial neural network. The 106 samples were randomly put into a training group (n=70,35 patients with liver cancer, 14 patients with hepatocirrhosis, and 21 healthy individuals) and a blind test group (n=36,17 patients with liver cancer, 8 patients with hepatocirrhosis, and 11 healthy individuals). Results The serum protein fingerprint pattern model obtained by the artificial neural network from training group was used to detect the 36 unknown serum samples. The sensitivity and specificity of this method in detection of liver cancer were 88.2%(15/17) and 94.6%(18/19)respectively. Conclusion In comparison with the traditional methods, this new method has higher sensitivity and specificity in diagnosis of liver cancer and should be further studied.

关 键 词：诊断 肝癌 蛋白质指纹图谱 血清蛋白质 健康人 AFP 肝硬化 应用 敏感性 利用 

分 类 号：R735.7[医药卫生—肿瘤]