Research on artificial intelligence-assisted diagnosis and treatment model of hemorrhagic strokes based on CT image data
Aims:This paper aims to Establish an artificial intelligence-assisted diagnosis and treatment model for hemorrhagic strokes.Methods:Based on the absolute and relative increase of hematoma volume,the occurrence of hematoma in patients was determined.Taking into account the clinical and the imaging information,the risk of hematoma expansion in patients was predicted using the random forest method.A subcategory grouping method for patient edema progression patterns based on Gaussian time series fitting was proposed;and the effects of different therapies on edema volume progression were obtained.Results:The accuracy of the hematoma predicting model reached 0.819;and the first two important treatment methods were intracranial pressure reduction and sedation and analgesia,which accounted for 23.4％and 22.1％of the effectiveness of all treatment methods.Conclusions:The hematoma prediction model can effectively predict the occurrence of hematoma in patients;and the cranial pressure reduction and sedation/analgesia treatment methods have a most important effect of edema control.
hemorrhagic strokerandom forest methodartificial intelligence-assisted diagnosis and treatment
万方数据
维普
