应用MALDI-TOF MS技术建立贲门癌血清蛋白指纹图谱诊断模型

Construction of gastric cardia adenocarcinoma diagnostic model using MALDI-TOF MS serum protein profiling

背景与目的:贲门癌是常见的消化道肿瘤,预后差,5年生存率低。早期诊断是改善其预后的关键因素之一。近年来,蛋白质组学的迅速发展推进了肿瘤标志物研究的进程,其中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)作为常用的蛋白质组学技术,成为检测和验证血液等体液肿瘤标志物的新技术及平台。本研究利用MALDI-TOF MS技术检测贲门癌患者血清蛋白指纹图谱,建立贲门癌诊断模型,探讨其临床应用价值。方法:收集河北医科大学第四医院2009年3月—2010年5月胸外科贲门癌患者血清63例和健康志愿者血清54例,采用弱阳离子蛋白芯片(WCX磁珠)对血清进行分析前处理,MALDI-TOF MS技术进行血清蛋白图谱检测,所得结果用ZUCI-蛋白芯片数据分析系统进行处理。运用遗传算法(genetic arithmetic,GA)结合支持向量机(support vector machine,SVM)运算建立贲门癌蛋白指纹图谱诊断模型:将117例标本随机分为训练组和盲法测试组,经训练后验证诊断模型的特异度和灵敏度。结果:采集贲门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69个有显著性差异的质荷比峰(P<0.01);从中筛选出差异最显著的10个蛋白质荷比峰建立诊断模型(m/z分别为2 863、4 655、3 883、3 241、3 952、2 295、2 741、1 546、4 054和2 671)。通过验证,该诊断模型的灵敏度为96.83%、特异度为90.74%。结论:应用MALDI-TOF MS技术能够检测贲门癌患者血清蛋白指纹图谱,建立贲门癌诊断模型。该模型在贲门癌诊断中具有较高的灵敏度和特异度,有一定的临床应用价值。

Background and purpose：Gastric cardia adenocarcinoma is a common digestive hact cancer with poor prognosis and a low 5-year survival rate. Early diagnosis is critical for improving prognosis. In recent years, fast development of proteomics has facilitated numerous attempts for cancer biomarker discovery. Matrix assisted laser desorption ionization time-of-flight mass spectrometry （MALDI-TOF MS）, one of the most widely used proteomic technologies, made it possible to analyze cancer biomarkers in the blood and other body fluid. This study aimed to examine the serum protein fingerprint model of GCA by MALDI-TOF MS, so as to set up a diagnostic model of GCA and investigate its clinical significance. Methods：Sixty three GCA patients and 54 healthy controls were obtained from Fourth Affiliated Hospital of Hebei Medical University during March of 2009 to May of 2010. Serum protein was extracted by the weak cation exchange （WCX） protein chip system, and protein fingerprint was examined using MALDI-TOF MS. Data was analyzed by ZUCI-protein chip data analyze system （ZUCI-PCDAS） and a GCA diagnostic model was established by genetic （GA） combined support vector machine （SVM）. One hundred and seventeen samples were randomly divided into training set and blinding test sets to examine the specificity and sensitivity of this diagnostic model. Results：Serum protein fingerprints of GCA patients and healthy controls were obtained by MALDI-TOF MS while the m/z （mass to charge） peaks of 69 differential proteins were obtained afterwards and analyzed using the ZUCI-PCDAS software package （P0.01）. From this process, 10 differential proteins （whose m/z peaks were 2 863, 4 655, 3 883, 3 241, 3 952, 2 295, 2 741, 1 546, 4 054, 2 671, respectively） were selected to establish a diagnostic model for cardia cancer. Specificity and sensitivity in this model for diagnosing cardia cancer was 96.07% and 92.42%, respectively. Conclusion：A GCA diagnostic model established from a serum protein fingerprint of GCA using MALDI-TOF MS and has shown high sensitivity and specificity for diagnosing cardia cancer.