基于无先兆偏头痛患者杏仁核异常有效连接预测非甾体抗炎药疗效的可行性  

Feasibility of predicting nonsteroidal anti-inflammatory drugs efficacy based on abnormal effective connectivity of amygdala in migraineurs without aura

作　　者：魏恒乐 张宏 陈季南 王锦锦 余玉盛 Wei Hengle;Zhang Hong;Chen Jinan;Wang Jinjin;Yu Yusheng(Department of Radiology,The Affiliated Jiangning Hospital of Nanjing Medical University,Nanjing City,Jiangsu Province 211100,China;Department of Neurology,The Affiliated Jiangning Hospital of Nanjing Medical University,Nanjing City,Jiangsu Province 211100,China)

机构地区：[1]南京医科大学附属江宁医院医学影像科,南京市211100 [2]南京医科大学附属江宁医院神经内科,南京市211100

出　　处：《中华疼痛学杂志》2022年第5期656-663,共8页Chinese Journal Of Painology

基　　金：南京市科技发展项目(YKK20202);南京医科大学科技发展项目(NMUB2020168)。

摘　　要：目的利用静息态功能磁共振成像技术探讨无先兆偏头痛(MwoA)患者杏仁核有效连接异常预测非甾体抗炎药(NSAIDs)疗效的可行性。方法收集2019年1月至2021年1月于南京医科大学附属江宁医院神经内科就诊并拟接受NSAIDs治疗的MwoA患者,年龄20~60岁,性别不限,均行静息态功能磁共振成像检查。依据药物疗效将患者分为有效组和无效组,应用格兰杰因果分析(GCA)方法计算组间两侧杏仁核异常有效连接,将其作为影像学特征,联合支持向量机(SVM)机器学习算法构建MwoA患者NSAIDs疗效的预测模型。采用受试者工作特征曲线(ROC)分析模型对NSAIDs的预测效能。结果最终70例患者纳入研究,有效组和无效组MwoA患者各35例。与无效组相比,有效组患者左侧杏仁核到右侧颞枕联合区(TOAC)和左侧枕中回有效连接减弱;反之,右侧TOAC和左侧枕中回到左侧杏仁核有效连接增强。并且,左侧丘脑到左侧杏仁核有效连接减低;右侧TOAC到右侧杏仁核有效连接增强。无效组患者的左侧杏仁核到右侧TOAC的有效连接与病程呈正相关(r=0.343,P=0.027)。基于SVM算法构建的预测模型ROC曲线下面积、敏感度和特异度分别为0.750、0.625和0.833。结论MwoA患者杏仁核与视觉相关皮层间的有效连接可预测NSAIDs的临床疗效。Objective To investigate the feasibility of predicting the efficacy of nonsteroidal anti-inflammatory drugs(NSAIDs)in patients with migraine without aura(MwoA)through abnormal amygdala-related effective connectivity using resting-state functional magnetic resonance imaging(MRI).Methods From January 2019 to January 2021,MwoA patients taken NSAIDs treatment were recruited in the Neurology Department of Nanjing Jiangning Hospital.All patients accepted brain MRI.According to the guidelines for prevention and treatment of migraine,MwoA patients were divided into effective group and non-effective group.Granger causality analysis was used to detect the abnormal effective connectivity between the bilateral amygdalae and whole brain.Then,combining the abnormal effective connectivity and machine learning algorithms,such as support vector machine(SVM),to construct the predictive model for NSAIDs efficacy in MwoA patients.Receiver operating characteristic curve(ROC)was performed to evaluate the prediction performance of models.Results A total of 70 patients were finally included in this study(35 cases in each group).Compared with non-effective group,effective group showed a weakened effective connectivity from left amygdala to right temporo-occipital association cortex(TOAC)and right middle occipital gyrus(MOG),while an enhanced effective connectivity from right TOAC and left MOG to left amygdalae.Moreover,a decreased effective connectivity from left thalamus to left amygdala and an increased effective connectivity from right TOAC to right amygdala were detected too.In addition,the effective connectivity from left amygdala to right TOAC was positively correlated with disease duration(r=0.343,P=0.027)in non-effective group.The area under ROC curve of SVM,sensitivity and specificity were 0.750,0.625 and 0.833,respectively.Conclusion Abnormal amygdala-related effective connectivity with visual-related cortex in MwoA patients could predict the clinical efficacy of NSAIDs.

关 键 词：消炎药 非甾类 格兰杰因果分析 机器学习算法 无先兆的偏头痛 

分 类 号：R747.2[医药卫生—神经病学与精神病学]