摘要
目的建立胃癌人工神经网络蛋白分子诊断模型,寻找胃癌早期诊断的新方法。方法采用表面增强激光解析离子化飞行时间质谱(SELDI-TOF-MS)技术及配套芯片CM10和Biomarker Wizard 3.1软件筛选胃癌差异表达蛋白,通过人工神经网络(ANN)建立并验证胃癌的SELDI分子诊断模型。结果共建立3个胃癌诊断模型(Ⅰ、Ⅱ、Ⅲ),分别为胃癌的诊断、筛查、鉴别诊断模型。选其中由5个差异表达蛋白(质荷比为2502、2544、3085、8574、8740)组成的胃癌人工神经网络诊断模型Ⅰ作为胃癌人工神经网络诊断模型,对胃癌的诊断灵敏度为95.0%,特异度为98.33%,阳性预测值为95.0%,阴性预测值为98.33%,诊断准确度为97.5%。结论 SELDI-TOF-MS技术对胃癌的早期诊断具有一定的价值,值得进一步的研究。
Objective To study serum protein of patients with gastric cancer and screen different expressed protein to find molecule diagnostic model and new method for early detection of gastric cancer.Methods The SELDI-TOF-MS,CM10 protein chip and Biomarker Wizard 3.1 software were performed to detect mass spectrogram of patients with gastric cancer and contrast persons for serum protein signature analysis to build and test ANN model.Results Three diognostic models were built for diagnosis /differentiatial diagnosis and screening of gastric cancer.Five specific protein peaks(M/Z as 2502,2544,3085,8574,8740) were chosen to develop the artificial neural network diagnostic model.The model was yielded a sensitivity of 95.0%,a specificity of 98.33%,a positive predictive value of 95.0%,a negative predictive value of 98.33%,a accuracy of 97.5%.Conclusion SELDI-TOF-MS is valuable to the early diagnosis of gastric cancer and need further study.
出处
《重庆医学》
CAS
CSCD
北大核心
2012年第1期22-24,共3页
Chongqing medicine
基金
国家高技术研究发展计划(863计划)重点基金资助项目(2006AA02090407)
关键词
胃肿瘤
神经网络模型
分子诊断模型
stomach neoplasms
neural networks
molecule diagnostic model