MALDI-TOF质谱筛查NSCLC患者血清特异性多肽的探索性研究

Preliminary Study of MALDI-TOF Mass Spectrometry-based Screening of Patients with the NSCLC Serum-Specific Peptides

背景与目的早期诊断是提高肺癌生存率的关键,传统的肺癌诊断技术仍存在一定局限性。鉴于近年来以质谱为核心技术的肿瘤蛋白组学在癌症诊断方面的初步研究,本研究探索性应用基质辅助激光解析电离飞行时间质谱(matrix assisted laser desorption ionization-time of ight-mass spectrometry,MALDI-TOF-MS)分析非小细胞肺癌(non-small cell lung cancer,NSCLC)患者和健康人群的血清差异多肽,以建立NSCLC的血清分类模型。方法将年龄和性别匹配的133例NSCLC患者和132例健康者血清标本按照3:1的比例随机分为两组:训练组由100例NSCLC患者和100例健康者血清标本组成,用以建立分类模型;测试组由33例NSCLC患者和32例健康者血清标本组成,用以验证模型。采用铜离子鳌合纳米磁珠提取血清多肽、MALDI-TOF-MS技术检测得到质谱图。ClinProToolsTM统计软件分析训练组NSCLC患者与健康者之间的多肽图谱,从中筛选出一组差异多肽并建立分类模型,最后用测试组对模型进行盲样验证。结果在训练组中观察到血清质荷比(m/z)在1,000Da-10,000Da范围内有131个差异多肽信号峰,在此范围内共得到14个有统计学意义的差异多肽峰(P<0.000,001;AUC0.9),其中NSCLC患者与健康者相比,表达上调的多肽有2个,表达下调的有12个,由统计软件筛选出3个多肽峰(7,478.59Da、2,271.44Da、4,468.38Da)建立分类模型,然后对测试组进行验证,其盲样验证敏感性100%,特异性96.9%,准确率98.5%。结论本组研究显示NSCLC患者与健康人群的血清多肽存在差异,应用MALDI-TOF-MS技术可建立NSCLC的血清多肽分类模型且小规模验证具有较好的敏感性和特异性,希望大规模验证模型,并与传统诊断方法对照或结合,进而尝试建立一种新的NSCLC早期诊断模式。

Background and objective The improved survival of patients with lung cancer depends on early diagnosis of lung cancer. However, the traditional diagnostic techniques have several limitations. Mass spectrometry （MS） has been applied as a core technology for cancer diagnosis in preliminary proteomic studies. The aim of this study is to explore the differences in the serum peptide levels of patients with non-small cell lung cancer （NSCLC） and healthy individuals using matrixassisted laser desorption/ionization （MALDI）-time-of-flight （TOF）-MS. A NSCLC serum classification model was then established. Methods One hundred and thirty three cases ofpatients with NSCLC serum specimens and 132 cases ofhealthy human serum specimens were randomly divided into two groups in accordance with the ratio of three to one without age and gender differences. The training group was used to establish the classification model, this group included serum samples from 100 NSCLC cases and 100 healthy individuals. The test group for validating the proposed model was composed of the remaining serum samples from 33 NSCLC cases and 32 healthy individuals. Peptides were extracted from the samples using magnetic beads - immobilized metal affinity capture - copper, and their mass spectra were obtained using an automated MALDI-TOFMS system, qqae MS data from the training group was analyzed using the ClinproToolTM software to identify the individual peptide fragments and establish the classification model. The sensitivity and specificity of the model were verified by blind testing with the test group. Results Among the 131 different peptide peaks, ranging from m/z 1,000 Da to 10,000 Da, 14 peaks were significantly different in the NSCLC samples of the training group, as compared with the controls （P〈0.000,001; AUC_〉 0.9）; these included 2 higher peaks and 12 lower peaks. The classification model was established, and the test group was verified for only 3 peptide peaks （7,478.59, 2,271.44 and 4,468.38 Da）, which were selected by the statistical software. Blind testing revealed that the proposed method had 100% sensitivity, 96.9% specificity and 98.5% accuracy. Conclusion Our results showed that the serum peptide levels were significantly different between NSCLC patients and healthy individuals. A serum peptide-based classification of NSCLC patients was established using an automated MALDI-TOF-MS system. This method demonstrated high sensitivity and specificity in a small-scale test. Future studies should test the proposed model through mass validation. The model could be compared or combined with traditional diagnostic methods to establish novel techniques for the early diagnosis of patients with NSCLC.