基于MALDI-TOF-MS技术建立非小细胞肺癌诊断预测模型及其初步验证

Establishment and preliminary validation of a diagnostic prediction model based on MALDI-TOF-MS for non-small cell lung cancer

目的应用基质辅助激光解析电离飞行时间质谱(MALDI-TOF-MS)技术结合ClinProTools系统软件筛选非小细胞肺癌(NSCLC)患者与体检健康者血清中的差异蛋白/多肽,建立NSCLC诊断预测模型,寻找NSCLC潜在血清肿瘤标志物。方法收集经组织病理学活检确诊的56例NSCLC患者及56例体检健康者血清标本,按照3∶1比例分为训练组(NSCLC患者42例,体检健康者42例)和验证组(NSCLC患者14例,体检健康者14例)。采用弱阳离子交换磁珠(WCX-MB)富集血清标本中的低丰度蛋白/多肽,应用MALDI-TOF-MS技术进行检测,得到相应的蛋白/多肽指纹图谱,筛选两组间差异蛋白/多肽,建立NSCLC诊断预测模型,并进行初步验证。结果经ClinProTools系统软件分析,筛选出11个多肽峰差异有统计学意义(P<0.000001)。与体检健康者相比,NSCLC患者质荷比为5906.73与2953.73的两个峰差异最明显,其受试者工作特征曲线下面积分别为0.96、0.86。运用遗传算法建立NSCLC诊断预测模型,经盲法验证,此模型灵敏度为92.9%,特异度为91.7%,准确度为92.3%。结论NSCLC患者与体检健康者血清中存在差异性蛋白/多肽,应用MALDI-TOF-MS技术建立的诊断预测模型能较好地用于NSCLC辅助诊断,为NSCLC早期诊断提供了一种新的策略。

Objective To screen the differential peptides in the serum of patients with non-small cell lung cancer(NSCLC)and healthy controls based on matrix-assisted laser desorption/ionization time-of-flight mass spectrometry(MALDI-TOF-MS)combined with ClinProTools system software,and to establish a diagnostic predictive model for finding potential tumor markers for NSCLC.Methods Serum samples from 56 patients with NSCLC diagnosed by pathology and 56 healthy controls were divided into training group(42 patients and 42 healthy controls)and validation group(14 patients and 14 healthy controls).The low-abundance proteins/peptides in serum samples was enriched by weak cation-exchange chromatography Magnetic Beads(WCX-MB),and the corresponding protein expression fingerprints were obtained by MALDI-TOF-MS.Differential proteins/peptides were screened and a diagnostic model for preliminary validation of NSCLC with ClinProTools system software was established.Results A total of 11 peptide peaks selected through ClinProTools system software had significant statistical differences(P<0.000001).The area under the receiver operating characteristics curve values of the two peaks m/z 5906.73 and m/z 2953.73 which were most significant were 0.96 and 0.86 respectively,both of which were up-regulated in the NSCLC patients.The validation results showed that the sensitivity,specificity and accuracy of the diagnostic prediction model established by the Genetic Algorithm were 92.9%,91.7%and 92.3%respectively.Conclusion There are differential proteins/peptides in the serum of NSCLC patients and healthy controls.The diagnostic model of NSCLC established by MALDI-TOF-MS could be used for NSCLC diagnosis,which provides a new strategy for early diagnosis of NSCLC.