机器学习在预测急性缺血性卒中预后中的应用进展OA

Acute ischemic stroke(AIS)is characterized by high mortality,high disability,and high recurrence rates.Its prognosis is affected by multiple interacting factors,including infarct location,collateral circulation,treatment time window,and baseline neurological function status.Traditional prognostic evaluation methods have limited ability in dealing with high-dimensional data and insufficient prediction accuracy,so it is difficult to meet the needs of accurate prediction.Machine learning(ML)can predict the poor prognosis of AIS more efficiently and accurately.This article summarizes the common ML algorithms,including traditional methods(such as support vector machine,decision tree,random forest,extreme gradient boost,ensemble learning,etc.)and deep learning methods(such as convolutional neural networks,etc.),as well as the application status and performance in AIS prognosis prediction.This article summarizes the advantages of multimodal data fusion modeling in improving the prediction accuracy,focusing on five aspects:functional outcome,treatment response,mortality risk,recurrence risk and complications.Compared with traditional models,the prediction models constructed by ML have superior performance and are expected to provide strong support for individualized management and clinical decision-making for AIS patients.