职业性三氯乙烯药疹样皮炎诊断模型的建立  

The establishment of diagnosis model on occupational medicamentosa-like dermatitis induced by trichloroethylene

作　　者：叶金波[1] 刘威[1] 周桂凤[1] 任晓虎[2] 黄培武[2] 洪文旭[2] 黄海燕[2] 刘建军[2] 

机构地区：[1]湖南师范大学医学院,长沙410006 [2]深圳市疾病预防控制中心现代毒理学重点实验室

出　　处：《中华预防医学杂志》2013年第2期151-154,共4页Chinese Journal of Preventive Medicine

基　　金：基金项目：国家自然科学基金（81273126,30972454）;深圳市重点实验室提升发展项目（cxB201005260068A）

摘　　要：目的应用弱阳离子交换磁珠（magneticbeadsbasedweakcationexchangechromatography，MB—WCX）、基质辅助激光解析电离飞行时间质谱（matrix—assistedlaserdesorptionionizationtime—of-flightmassspectrometry，MALDI—TOF—MS）和ClinProTools生物信息学方法检测职业性三氯乙烯药疹样皮炎（occupationalmedicamentosa—likedermatitisinducedbytrichlor（1ethylene，OMLDT）患者的血清多肽指纹图谱，建立OMLDT诊断模型。方法收集2009年12月至2010年10月经深圳市职业病防治院诊断的OMLDT患者和对照人群血清样品各28份，选取其中的患者和对照人群血清样品各14份作为建模组，采用MB—WCX联合MALDI—TOF—MS技术检测血清多肽指纹图谱，筛选OMLDT特征性多肽标志并建立OMLDT疾病蛋白质组学诊断模型。用其余的14份患者和14份对照人群的血清样作为验证组，评价模型的准确度和识别率。结果应用ClinProTools软件，共得到159个峰，33个为有统计学意义的峰（P〈0．05）。其中，相对于对照组，建模组的病例组中有20个峰表达降低，有13个表达增高。采用监督神经网络算法（supervisedneuralnetworkalgorithm，SNN）对多肽峰进行筛选，质荷比（m／z）为2106．29和3263．78的2个多肽峰受试者工作特征曲线（receiveroperatingcharacteristiccurve，ROC）的下面积（TheareaundertheROCcurve，AUC）最接近1，最能区分病例组和对照组，2D分布图上也能明显区分，选择这2个多肽峰构建OMLDT的诊断模型。诊断模型的交叉验证和识别能力分别是87．5％和98．5％，灵敏度和特异度分别为84．8％和82．1％。结论应用MB—WCX、MALDI—TOF—MS技术结合ClinProTools软件首次对OMLDT建立诊断模型并验证，筛选到了特异性差异多肽，具备较高的灵敏度和特异度，为临床早期诊断提供了科学依据。Objective Based on magnetic beads based weak cation exchange chromatography （MB- WCX） , matrix assisted laser desorption ionization time of flight mass spectrometry （MALDI-TOF-MS） and ClinProTools software, the polypeptides of serum about occupational medicamentosa-like dermatitis induced by trichloroethylene （OMLDT） patients were studied, and a diagnostic model of OMLDT was built. Methods According to diagnostic criteria of OMLDT, serum of 28 OMLDT patients and 28 controls which were diagnosed by Shenzhen prevention and treatment center for occupational disease were collected. With the combination of MB-WCX and MALDI-TOF-MS,the polypeptides fingerprint of serum of 14 OMLDT patients and 14 controls were detected, what＇s more, the ClinProTools software and SNN algorithm was used for screening characteristic polypeptides and establishing diagnostic model of OMLDT. Then other objects were applied to validate the model to evaluate accuracy. Results A total of 159 peaks were attained by ClinProTools software, of which 33 peaks were statistical content （ P 〈 0. 05 ）. What is more, comparing with the control group ,20 peaks in case group were decreased, and 13 peaks were increased. Two peaks of them were identified, that is 2106. 29 and 3263.78, to classify and determine that two groups by receiver operating characteristic curve（ROC） analysis. 2D peaks distribution map certified this finding and the area under the ROC curve was closed to 1. A model was established by SNN algorithm, whose cross validation and recognition capability were 87.5% and 98.5% ,respectively. Its sensitivity and specificity were 84. 8% and 82. 1%, separately, which displayed good separating capacity. Conclusion In the combination of MB- WCX, MALDI-TOF-MS and ClinProTools software, specifical different polypeptides were screened and OMLDT diagnostic model was built primarily. Also, the model and the results were positively validated, which would play a significant role in early diagnosis.

关 键 词：三氯乙烯 皮炎 职业性 诊断 血清多肽指纹图谱 

分 类 号：R135.7[医药卫生—劳动卫生]