论文部分内容阅读
目的检测甲状腺癌、甲状腺良性结节和正常人血清中蛋白质组图谱,筛选特异的蛋白质标记物,构建用于甲状腺癌早期诊断的血清蛋白质指纹图谱模型。方法应用表面增强激光解吸电离飞行时间质谱(SELDI-TOF-MS)技术检测108例血清标本的蛋白质质谱,其中甲状腺癌40例(Ⅰ～Ⅱ期26例,Ⅲ～Ⅳ期14例),甲状腺良性结节36例,正常人32例。随机抽取87例标本(甲状腺癌32例,甲状腺良性结节30例,正常人25例)作为训练组,应用支持向量机和判别分析的方法分析质谱数据,建立甲状腺癌诊断模型,留一法交叉验证。结果区分甲状腺癌和正常人的诊断模型交叉检验敏感性87．5%,特异性80．0%,用15例未知血清盲法测试敏感性100%,特异性86%。区分甲状腺癌和甲状腺良性结节的诊断模型交叉检验敏感性81%,特异性87%,用14例未知血清盲法测试敏感性88%,特异性83%。结论SELDI-TOF-MS技术结合生物信息学方法检测甲状腺癌血清蛋白质指纹图谱是早期诊断甲状腺癌的一种特异性强,敏感性高的新方法,值得进一步研究。
Objective To detect the proteome map of thyroid carcinoma, benign thyroid nodules and normal human serum and screen specific protein markers to construct a serum protein fingerprint model for the early diagnosis of thyroid cancer. Methods Serum samples from 108 cases were detected by SELDI-TOF-MS, including 40 thyroid carcinomas (26 cases in stage Ⅰ ~ Ⅱ, 14 cases in stage Ⅲ ~ Ⅳ), benign thyroid Nodules in 36 cases, 32 cases of normal. A total of 87 specimens (thyroid cancer 32, thyroid benign nodules 30, and normal 25) were randomly selected as the training group. Mass spectrometry data were analyzed by support vector machine and discriminant analysis to establish a diagnostic model of thyroid cancer, verification. Results The diagnostic sensitivity of differentiated thyroid cancer from that of normal controls was 87.5% and 80.0%, respectively. Sensitivity and specificity were 100% and 86%, respectively, with 15 unknown serum blind tests. The diagnostic test that differentiated between thyroid cancer and benign thyroid nodules had a sensitivity of 81% and a specificity of 87%, with a sensitivity of 88% and a specificity of 83% using a blinded test of 14 unknown sera. Conclusion SELDI-TOF-MS combined with bioinformatics methods to detect serum protein fingerprinting of thyroid cancer is a new and specific method for early diagnosis of thyroid cancer. It is worth further study.