APP下载

血清蛋白指纹图谱检测对儿童支原体肺炎早期诊断的研究

2017-07-01严红梅阮琰林伟唐伟伟熊丽君黄明

中外医疗 2017年13期

严红梅+阮琰+林伟+唐伟伟+熊丽君+黄明翔+翁丽珍

[摘要] 目的 用表面增強激光解析电离飞行时间质谱(SELDI-TOF-MS)技术和蛋白质芯片技术检测儿童支原体肺炎(MPP)和非MPP患者血清蛋白指纹图谱,探讨基于人工神经网络的的蛋白指纹图谱诊断模型对儿童MPP早期诊断的价值。方法 用SELDI-TOF-MS技术和蛋白质芯片技术测定2014年11月—2016年6月该院儿科住院的130例血清标本的蛋白指纹图谱,其中MPP 68例,非MPP 62例。将测定的血清蛋白指纹图谱建立人工神经网络预测模型,用随机抽取的75例标本(MPP 40例,非MPP 35例)作为训练组,进行训练与交叉验证,另外55例标本(MPP 28例,非MPP 27例)作为测试组,进行盲法验证该模型。结果 利用从75例训练组得出的基于人工神经网络的由M/Z为4 094.91、4 650.51、5 825.38、6 031.77、7 862.79、10 122.04组成的MPP血清蛋白指纹图谱决策树诊断模型,对MPP诊断的敏感性95.00%(38/40),特异性为88.57%(31/35),阳性率为92.00%(69/75),ROC曲线下面积为0.975,利用测试组的55例标本对模型进行验证,结果显示,对MPP诊断的敏感性89.29%(25/28),特异性为96.30%(26/27),阳性率为92.27%(51/55)。结论 该研究建立的支原体肺炎蛋白指纹图谱诊断模型对支原体肺炎的检测时间短,诊断准确率高,有助于诊断支原体肺炎。

[关键词] 蛋白指纹图谱技术;实验室诊断技术;儿童支原体肺炎

[中图分类号] R5 [文献标识码] A [文章编号] 1674-0742(2017)05(a)-0004-05

[Abstract] Objective Enhanced laser desorption ionization time-of-flight mass spectrometry(SELDI-TOF-MS) with the purpose of detecting mycoplasma pneumonia in children and the technology of protein chip technology(MPP) and non MPP serum protein fingerprint, explore the early diagnosis of children MPP protein fingerprints of the artificial neural network diagnostic model based on the value of. Methods The protein fingerprints of 130 serum samples which be in pediatrics from Nov. 2014 to Jun. 2016 were determined by SELDI-TOF-MS and protein chip technique, including MPP 68 cases and non MPP 62 cases. The determination of serum protein fingerprint to establish artificial neural network prediction model, using 75 samples randomly selected (MPP 40 cases, MPP 35 cases) as the training group, training and cross validation, the other 55 specimens (MPP 28 cases, MPP 27 cases) as the test group, the blind method to verify the model. Results From 75 cases of training group based on the artificial neural network by M/Z was 4 094.91, 4 650.51, 5 825.38, 6 031.77, 7 862.79, 10 122.04 OMPP serum protein fingerprint decision tree diagnosis model, the sensitivity of 95.00% in diagnosis of MPP (38/40), the specificity was 88.57% (31 /35), the positive rate was 92.00% (69/75), the area under the ROC curve was 0.975, verified the model by using 55 samples test Results showed that the sensitivity of 89.29% in diagnosis of MPP(25/28), the specificity was 96.30% (26/27), the positive rate was 92.27% (51/55). Conclusion The diagnostic model of Mycoplasma pneumoniae protein fingerprint established in this study can be used for the diagnosis of Mycoplasma pneumonia.

[Key words] Protein fingerprinting technology; Laboratory diagnosis; Mycoplasma pneumonia in children

有研究显示病原体为肺炎支原体(Mycoplasma,MP)的支原体肺炎(Mycoplasma pneumoniae pneumonia,MPP),在因社区获得性肺炎(Community acquired pneumonia,CAP)住院的儿童患者中所占比例可达10%~40%[1-2]。传统认为MPP大多数临床症状相对较轻,经大环内酯类抗生素治疗后症状可好转,且支原体感染多有自限性,预后良好。但是近几年,临床表现为难治性支原体肺炎(RMPP)及重症支原体肺炎(SMPP)的患者逐渐增多[3]。……

登录APP查看全文