影像诊断与介入放射学

目的 探讨磁共振各向同性容积扫描(CUBE)序列对成人病毒性脑炎继发癫痫诊断的增益价值。方法收集2022年1月—2024年1月248例成人病毒性脑炎患者临床资料进行回顾性研究，根据是否继发癫痫分为癫痫组和非癫痫组，均在1.5 T磁共振上完成常规平扫各序列及矢状位增强CUBE T₂液体衰减反转恢复(FLAIR)序列扫描。统计两组一般资料、病变与正常灰白质及脑积液的对比率，采用Logistic回归分析对比率与成人病毒性脑炎继发癫痫的关系，采用受试者工作特征(ROC)曲线分析对比率预测成人病毒性脑炎继发癫痫的价值，DeLong检验曲线下面积(AUC)大小，并进行外部验证。结果 癫痫组患者脑脊液压力及血清神经元特异性烯醇化酶(NSE)、tau蛋白表达、病变/白质对比率、病变/灰质对比率及脑脊液单纯疱疹病毒(HSV)(+)占比均高于非癫痫组(P<0.05)。校正脑脊液压力、脑脊液HSV(+)、血清NSE及tau蛋白后，病变/白质对比率、病变/灰质对比率仍是成人病毒性脑炎继发癫痫的影响因素(P<0.05)。根据多因素Logistic回归分析结果构建模型1[脑脊液压力、脑脊液HSV(+)、NSE、磷酸化tau(p-tau)蛋白]和模型2[脑脊液压力、脑脊液HSV(+)、NSE、p-tau蛋白、病变/白质对比率、病变/灰质对比率]，并绘制病变/白质对比率、病变/灰质对比率、模型1、模型2的ROC曲线，结果显示病变/白质对比率、病变/灰质对比率预测成人病毒性脑炎继发癫痫的AUC为0.763、0.768；模型2预测成人病毒性脑炎继发癫痫的AUC为0.928，高于模型1的AUC 0.886(P<0.05)。外部验证显示，模型2预测成人病毒性脑炎继发癫痫的准确度(95.71%)高于模型1(81.43%)(P<0.05)。结论 磁共振CUBE序列在成人病毒性脑炎继发癫痫预测中有增益价值，有助于优化临床治疗决策。

Objective To explore the incremental value of isotropic volumetric scanning(CUBE) in adults with epilepsy secondary to viral encephalitis. Methods A retrospective study was conducted on the clinical data of 248 adult patients with viral encephalitis from January 2022 to January 2024. Patients were divided into two groups with or without secondary epilepsy. All patients underwent conventional MRI and sagittal contrast-enhanced CUBE T₂ fluid-attenuated inversion recovery(FLAIR)-weighted scanning on a 1.5 T magnetic resonance scanner. The general clinical information and the lesion-to-normal gray matter, white matter, and cerebrospinal fluid(CSF) contrast ratios were analyzed for both groups. Logistic regression analysis was used to explore the correlation between the contrast ratios and secondary epilepsy. The contrast ratio for predicting secondary epilepsy was assessed using a receiver operating characteristic(ROC) curve. The area under the ROC curve(AUC) was calculated using the DeLong test. External validation was also performed. Results The CSF pressure, serum neuron-specific enolase(NSE), tau protein expression, lesion/white matter contrast ratio, lesion/gray matter contrast ratio, and the proportion of herpes simplex virus in CSF(HSV+) in the epilepsy group were all significantly higher than those in the non-epilepsy group(P <0.05). After adjusting for CSF pressure, CSF HSV+, serum NSE, and tau protein, the lesion/white matter and lesion/gray matter contrast ratios remained as influencing factors for viral encephalitis-induced epilepsy(P <0.05). Based on the results of the multivariate logistic regression analysis, model 1(CSF pressure, CSF HSV+, NSE, phosphorylated tau protein) and model 2(CSF pressure, CSF HSV+, NSE, phosphorylated tau protein, lesion/white matter contrast ratio, and lesion/gray matter contrast ratio) were constructed. The ROC curve analysis showed significantly(P <0.05) higher AUC of model 2(0.928) than those of lesion/white matter contrast ratio(0.763), lesion/gray matter contrast ratio(0.768), and model 1(0.886) for predicting viral encephalitis-induced epilepsy. External validation showed that model2 had also significantly(P <0.05) higher accuracy(95.71%) than model 1(81.43%). Conclusion The MRI CUBE sequence has incremental value for predicting secondary epilepsy and can aid management of viral encephalitis.