Predictive Value of CT Texture Analysis Combined with Machine Learning in Hemorrhagic Transformation of Acute Cerebral Infarction
Objective To explore the predictive value of CT texture analysis combined with machine learning in hemorrhagic transformation after acute cerebral infarction.Methods The CT image data of patients with acute cerebral infarction admitted to hospital from January,2021 to September,2023 were retrospectively analyzed,and the differences of morphological parameters of CT infarcted area between hemorrhagic transformation group(n=78)and non-hemorrhagic group(n=122)after thrombolytic therapy were compared.The texture feature parameters of infarcted area were extracted from CT images,and various machine learning models were constructed based on the texture and morphological features,respectively.Results The morphological characteristics of the infarct area and whether it was multiple infarcts were statistically significant(P＜0.05).The machine learning model based on texture features can better predict hemorrhagic transformation,and its overall efficiency is higher than that of morphological feature models,among which XGBoost and CatBoost have the highest prediction efficiency.Conclusion Texture analysis based on CT can effectively predict hemorrhagic transformation of acute cerebral infarction.
Cerebral InfarctionTexture AnalysisCT ImagingHemorrhagic TransformationPredicted Value
NETL
NSTL
万方数据
