基于Landmark模型动态预测老年人轻度认知障碍向阿尔茨海默病的转化

Dynamic Prediction of Conversion from Mild Cognitive Impairment to Alzheimer′s Disease based on Landmark Model

目的利用Landmark模型对轻度认知障碍(mild cognitive impairment,MCI)的老年人转为阿尔茨海默病(Alzheimer′s disease,AD)的概率进行动态估计,为早期发现高危AD患者提供帮助。方法利用312名MCI个体的纵向和生存数据构建三个landmark模型(模型1、模型2和模型3)。利用Brier得分和C指数评估模型的预测性能并选出最优模型进行动态预测。结果模型3的预测性能较好,且FAQ、RAVLT-immediate和海马体体积是MCI转为AD重要的预测变量。在不同随访年,利用模型3和这三个预测变量预测两名MCI个体两年后转为AD的概率。MCI个体1转为AD的概率逐年下降,属于AD低危个体;而MCI个体2转为AD的概率逐年上升,属于AD高危个体。结论本研究对MCI个体向AD转化的概率进行动态估计,可识别AD高危群体。

Objective To predict chances of conversion from mild cognitive impairment(MCI)to Alzheimer′s disease(AD)dynamically,and provide a methodological reference for early detection of high-risk AD patients.Methods We established three landmark models(model 1,model 2 and model 3)based on longitudinal and survival data of 312 MCI individuals and evaluated models by C-index and Brier score.The better model was used to predict conversion dynamically.Results The model 3 had better predictive performance and RAVLT-immediate,FAQ and hippocampal volume were significant predictors for conversion from MCI to AD.We predicted the 2-year chances of conversion to AD among two MCI individuals with model 3 and three significant predictors.The chances of conversion to AD of MCI individual 1 were decreasing annually,and he was at low risk of AD.While the chances of conversion to AD of MCI individual 2 were increasing annually,and he was at high risk of AD.Conclusion Our study predicted chances of conversion from MCI to AD and was helpful to identify high-risk AD individuals.