Radiomics Model Validates Cognitive Impairment Prediction in CSVD

A radiomics model using hippocampal texture was developed to predict different levels of cognitive impairment in cerebral small vessel disease (CSVD). The model's performance varied across groups. In distinguishing mild cognitive impairment (mild-CSVD) from normal controls, the model achieved an accuracy of 0.625, sensitivity of 0.828, and specificity of 0.316. For distinguishing mild-CSVD from severe cases (severe-CSVD), it reached an accuracy of 0.683, sensitivity of 0.167, and specificity of 0.897. Finally, when differentiating severe-CSVD from normal controls, the model exhibited an accuracy of 0.781, sensitivity of 0.538, and specificity of 0.947. These results indicate that the radiomics model based on hippocampal texture has potential but varied efficacy in diagnosing cognitive impairment levels in CSVD patients.