基于MALDI-TOF MS技术食管鳞癌诊断预测模型的建立及初步验证

Establishment and preliminary validation of a diagnostic prediction model for esophageal squamous cell carcinoma based on MALDI-TOF MS

目的建立食管鳞癌诊断预测模型,寻找食管鳞癌潜在肿瘤标志物。方法收集59例食管鳞癌患者及57例体检健康者血清标本,按照3∶1随机分为训练组(患者44例,健康人对照42例)和验证组(患者15例,健康人对照15例)。弱阳离子交换磁珠(WCX-MB)提纯血清蛋白质/肽后,经基质辅助激光解吸电离飞行时间质谱(MALDI-TOF MS)检测,筛选出差异表达蛋白质/肽,构建诊断预测模型并初步验证。结果经Clin ProTools软件分析,共筛选出差异多肽峰31个(P<0.05),其中18个峰差异有统计学意义(P<0.01);与健康人对照组相比,8个峰在食管鳞癌组中表达上调; 10个峰在食管鳞癌组中表达下调。其中差异峰m/z 2 660.84和m/z 5 336.49的ROC曲线下面积(AUCROC)分别为0.95和0.91,且均在食管鳞癌组中表达上调。应用遗传算法(GA)建立的食管鳞癌诊断预测模型经验证其准确性为93.10%,敏感性为92.90%,特异性为93.30%。结论建立的食管鳞癌诊断预测模型能较好的用于食管鳞癌辅助诊断,多肽峰m/z 2 660.84和m/z 5 336.49有望成为食管鳞癌潜在肿瘤标志物。

Objective To establish a diagnostic prediction model for esophageal squamous cell carcinoma(ESCC)and search the potential biomarkers of ESCC.Methods Serum samples from 59 patients with ESCC and 57 healthy controls were collected,and randomly divided into the training group(44 patients and 42 healthy controls)and validation group(15 patients and 15 healthy controls).Serum proteins/peptides were extracted and purified with weak cation-exchange chromatography Magnetic Beads(WCX-MB),and detected by the matrix-assisted laser desorption/ionization time-of-flight mass spectrometry(MALDI-TOF MS).Then the differentially expressed proteins/peptides were screened out,and a diagnostic prediction model for ESCC was established and preliminarily validated.Results The Clin ProTools software identified 31 differential peptide peaks(P<0.05),among which 18 peaks had significant difference(P<0.01).Compared with healthy controls,8 peaks were up-regulated in ESCC patients,while 10 peaks were down-regulated.Among them,the areas under the receiver operating characteristics(ROC)curve(AUCROC)of m/z 2 660.84 and m/z 5 336.49 peaks were 0.95 and 0.91,respectively,and their expressions were up-regulated in ESCC patients.The validation results showed that the accuracy,sensitivity and specificity of the diagnostic prediction model established by the genetic algorithm(GA)were 93.10%,92.90%and 93.30%,respectively.Conclusion The established diagnostic prediction model may be used for the auxiliary diagnosis of ESCC.Two peptide peaks of m/z 2 660.84 and m/z 5 336.49 may be the potential biomarkers of ESCC.