基于因果推断肺癌患者生存时间预测方法

PREDICTION METHOD OF SURVIVAL TIME OF LUNG CANCER PATIENTS BASED ON CAUSAL INFERENCE

医学上常用回归的方法评估肿瘤患者的生存时间,但有一定的局限性。为了提高回归结果,提出一种基于因果推断的患者生存时间预测方法(MRCI-DNN)。采用因果推断算法构建病理因素与患者生存时间的因果网络结构图,从因果网络结构图中筛选主要因素,并结合深度神经网络模型预测生存时间。实验表明,肺癌分期、放化疗、吸烟、PLR、肺癌类型及NLR是影响肺癌患者生存时间的主要因素。通过实验对比,基于因果推断方法筛选主要因素应用在深度神经网络预测上要优于其他选择特征方法。

Regression is commonly used in medicine to assess survival timeof cancer patients,but regression analysis exists certain limitations.This paper proposes a method for predicting patient survival time based on causal inference to improve the regression results.This method used a causal inference algorithm to construct a causal network structure diagram of pathological factors and patient survival time,and filtered the main factors from this diagram.DNN was combined to predict the survival time.The experimental results show that the main factors affecting survival time of lung cancer patients include lung cancer stage,chemotherapy,smoking,PLR,lung cancer type,and NLR.Through experimental comparison,this method outperforms other feature selection methods in the prediction of DNN.