Study on cognitive impairment in patients with cerebral small vessel disease based on MRI quantitative brain segmen-tation

LIANG Jiaxin; CHANG Zhuoran; FAN Zheng; ZHANG Zhenxian; LI Yanjing; HUANG Xiaoqi; ZHANG Jianbing

doi:10.20258/j.cnki.1006-9011.2025.11.003

Abstract：Objective To explore the diagnostic value of whole brain quantitative segmentation based on artificial intelli-gence software MRI and brain region volume in early cognitive impairment of patients with small cerebral vascular disease.Methods A total of 135 patients with CSVD were selected,including 71 with cognitive impairment and 64 without cognitive im-pairment.The U-AI artificial intelligence software was used for quantitative brain segmentation to analyze the differences in brain region volumes between the two groups,and the ROC curve was drawn to analyze the diagnostic value of each brain region vol-ume for cognitive impairment in CSVD patients.Results The relative volumes of the cerebrospinal fluid,periventricular cere-brospinal fluid,third ventricle,choroid plexus and lateral ventricle brain regions in the cognitive impairment group were larger than those in the control group,while,the relative volumes of the gray matter,white matter,hippocampus,thalamus,precen-tral gyrus and superior frontal gyrus brain regions were smaller than those in the control group,with statistically significant dif-ferences(all P<0.05).The relative volumes of the cerebrospinal fluid,periventricular cerebrospinal fluid,hippocampus,white matter and lateral ventricle brain regions were independent influencing factors for cognitive impairment in CSVD patients(all P<0.05).The relative volume of the hippocampus had the highest diagnostic value for early cognitive impairment in CSVD patients(AUC=0.891),with sensitivity of 70.3%and specificity of 84.5%.Conclusion Quantitative brain volume by MRI has certain diagnostic value for early cognitive impairment in CSVD patients,with the diagnostic efficacy of hippocampus volume being the best.