基于图像自动分割算法的MRI成像对脑小血管病患者脑白质病变的诊断分析  

Diagnosis of White Matter Lesions in Patients with Cerebrovascular Disease by MRI Based on Automatic Image Segmentation Algorithm

作　　者：党丽丽 刘瑞[2] DANG Li-li;LIU Rui(Department of Neurology,Yulin Fourth Hospital,Yulin 7190001,Shaanxi Province,China;Department of Neurology,Yulin First Hospital,Yulin 7190002,Shaanxi Province,China)

机构地区：[1]榆林市第四医院神经内科,陕西榆林719000 [2]榆林市第一医院神经内科,陕西榆林719000

出　　处：《中国CT和MRI杂志》2023年第9期31-32,41,共3页Chinese Journal of CT and MRI

基　　金：陕西省自然科学基金项目(2020JM-706)。

摘　　要：目的分析基于图像自动分割算法的磁共振扫描(MRI)成像对脑小血管病患者脑部病变的诊断价值。方法回顾性分析本院2021年1月~2022年4月收治的106例可疑脑小血管病患者的临床资料,构建基于图像自动分割算法的MRI成像方法对脑白质病变情况进行诊断。比较常规MRI和基于图像自动分割算法的MRI成像对脑白质病变的诊断的效能。结果106例受试者中有73例确诊为脑小血管病,余33例中有17例帕金森病、10例周围神经病、6例特发性震颤;脑白质病变常规MRI检查可见T1加权成像(T1WI)低信号,T2加权成像(T2WI)和液体衰减反转恢复序列(FLAIR)呈高信号,且可累及脑室周围、深部脑白质,早期可见散在、斑片状、大小不等病灶,晚期可见片状信号;基于图像自动分割算法的MRI成像检出脑白质病变诊断脑小血管病的灵敏度、特异度、准确度分别为89.04%、93.94%、90.57%,均高于常规MRI的68.49%、66.67%、67.92%(P<0.05),且前者的Kappa值为0.815,也高于后者的0.648。结论 基于图像自动分割算法的MRI成像可用于判断脑白质病变进而诊断脑小血管病。Objective To analyze the diagnostic value of magnetic resonance imaging(MRI)based on automatic image segmentation algorithm for brain lesions in patients with cerebrovascular disease.Methods The clinical data of 106 patients with suspected cerebrovascular disease admitted to our hospital from January 2021 to April 2022 were analyzed retrospectively,and an MRI imaging method based on image automatic segmentation algorithm was constructed to diagnose the white matter lesions.The diagnostic efficacy of conventional MRI and MRI based on automatic image segmentation algorithm for white matter lesions was compared.Results Among 106 subjects,73 cases were diagnosed as cerebrovascular disease,and the remaining 33 cases included 17 cases of Parkinson's disease,10 cases of peripheral neuropathy,and 6 cases of idiopathic tremor.Conventional MRI examination of cerebral white matter lesions showed low signal intensity on T1 weighted imaging(T1WI),high signal intensity on T2 weighted imaging(T2WI)and fluid attenuated inversion recovery sequence(FLAIR),and white matter around ventricle and deep brain could be involved.Scattered,patchy and unequal sized lesions could be seen in early stage,and patchy signals could be seen in late stage.The sensitivity,specificity and accuracy of MRI imaging based on automatic image segmentation algorithm in detecting cerebral white matter lesions in diagnosing cerebrovascular diseases were 89.04%,93.94%and 90.57%respectively,which were higher than those of 68.49%,66.67%and 67.92%of conventional MRI(P<0.05),and the Kappa value of the former was 0.815,which was also higher than 0.648 of the latter.Conclusion MRI imaging based on automatic image segmentation algorithm can be used to judge white matter lesions and then diagnose cerebrovascular diseases.

关 键 词：图像自动分割算法 磁共振扫描 脑小血管病 脑白质病变 

分 类 号：R445.2[医药卫生—影像医学与核医学]