摘要
目的寻找特异的蛋白质标记物，探讨神经母细胞瘤血清蛋白质标记物的检测及初步诊断模型的构建和临床应用。方法收集血清样本87例，其中47例为神经母细胞瘤患儿，30例为其它恶性实体肿瘤患儿，10例为健康儿童；用ZUCI—Protein Chip Data Analyze System分析软件进行数据处理；经留一法交叉验证，分类器评价模型的预测效果。结果构建3个模型并筛选出10个蛋白质标记物，能成功区分神经母细胞瘤和健康儿童蛋白质谱差异表达模型的敏感性为100．00％，特异性为100．00％，区分神经母细胞瘤术前和术后蛋白质谱差异表达模型的敏感性为100．00％，特异性为100．00％，区分神经母细胞瘤与其它恶性实体肿瘤血清蛋白质指纹图谱模型的敏感性为88．89％，特异性为100．00％。结论用SELDI—TOF—MS及生物信息学技术，并结合支持向量机（SVM）初步建立的模型可作为神经母细胞瘤的另一种特异性强、敏感性高的辅助检查手段。
Objective To search the specific protein biomarkers constructing the serum protein markers in the early dignosis of Neuroblastoma. Methods Eighty-seven serum samples （ including 47 Neuroblastoma patients,30 children malignant solid tumor 10 healthy children）were detected by SELDI-TOF-MS,and proceeded by ZUCI-protein chip Data Analyze System Software to be identified by Jackknife, and evaluate the model＇ s predictive effect. Results To analyze all the serum protein icon by cluster analysis can find a proteomies model. The detective model in combined with 10 biomarkers could separate neuroblastoma from the healthy group with sensitivity of 100. 00% , specificity of 100.00% ;separate preoperative neuroblastoma from postoperative neuroblastoma with sensitivity of 100.00%, specificity of 100.00% ;The diagnosis model combined with 5 biomarkers could separate neuroblastoma from other children malignant solid tumors with sensitivity of 88. 89% ,specificity of 100. 00%. Conclusion This detective model is a more effective method for the early diag- nosis neuroblastoma, and can develope to a new path to research for new tumor markers to study the proteomics in neuroblastoma. The model of neuroblastoma combinated by SELDI-TOF-MS with SVM and bioinformation technique can regard as another more specificity and sensitivity auxiliary examination mean.
出处
《临床小儿外科杂志》
CAS
2011年第5期329-332,336,共5页
Journal of Clinical Pediatric Surgery
基金
本研究为国家自然科学基金资助项目（项目号30571930）
