摘要
目的探讨人工神经网络(ANN)技术对肺癌的诊断价值。方法采用电化学发光免疫法分别测定胸腔积液及血清中肿瘤标志物癌胚抗原(CEA)、糖类抗原125(CA-125)、糖类抗原19-9(CA-19-9)和肿瘤特异性生长因子(TSGF)的水平,建立肿瘤标志物ANN模型,并验证该ANN模型对肺癌与肺良性疾病的鉴别诊断价值。结果4种肿瘤标志物联合检测的灵敏度为97.4%,特异度为56.1%,准确率为84.9%;ANN模型对肺癌鉴别诊断的灵敏度为100%,特异度为93.3%准确率为97.8%。结论ANN模型能够对肺癌和肺良性疾病进行鉴别诊断,可为肺癌提供临床辅助诊断。
Objective To investigate the application of artificial neural network (ANN) combined with 4 pleural fluid and serum tumor markers in the diagnosis of lung cancer. Methods The levels of the 4 tumor markers, CEA, CA 125, CA 19-9 and tumor specific growth factor(TSGF) were detected by electrochemiluminescene immunoassay, and the ANN model were established combined with 4 tumor markers, to evaluate the value of differential diagnosis for lung cancer and lung benign disease. Results The sensitivity,specificity and accurate rates of the combination of 4 tumor markers were 97.4%, 56. 1% and 84. 9%, respectively. The sensitivity, specificity and accurate rates of ANN model were 100%, 93.3 % and 97. 8%. respectively. Conclusion The ANN model can be used to discriminate between lung cancer and lung benign disease, and it can provide clinical assistant diagnosis for lung cancer.
出处
《山东医药》
CAS
北大核心
2008年第21期11-13,共3页
Shandong Medical Journal
基金
国家自然科学基金(30571552)
河南省中青年骨干教师资助项目
关键词
人工神经网络
肺肿瘤
癌胚抗原
糖类抗原
肿瘤特异性生长因子
artificial neural network
lung neoplasm
carcinoembryonic antigen
carbohydrate antigen
tumor specific growth factor