模式识别及人工神经网络技术在肺癌组织分型中的应用研究

Study of Histological Type of Lung Cancer by Pattern Recognition and Artificial Neural Network

目的:探讨模式识别及人工神经网络技术在肺癌组织分型中的应用。方法:用放射性免疫法测定了肺癌患者血清中4种肿瘤标志物(CEA、CA125、胃泌素及NSE)的水平,在此基础上采用模式识别及人工神经网络技术,探讨它们在肺癌组织分型中的应用价值。结果:在判别小细胞肺癌与非小细胞肺癌类型中,这些方法的总正确率均在85％以上。结论:模式识别及人工神经网络技术在肺癌组织分型中有一定的参考价值,同时为临床提供必要的参考资料。

Aim: To study histological type of lung cancer by pattern recognition and artificial neural network. Methods: The levels of four tumor markers were detected by radioimmunoassay. Pattern recognition and artificial neural network were applied to distinguish small cell from non-small cell lung cancer. Results: The total accuracy was above 85% in distinguishing SCLC from NSCLC. Conclusions: They were useful methods in histological type of lung cancer and could provide essential reference value for clinic.