基于统计参数图的脑内不对称分析法在^18F-FDG PET代谢显像对单侧颞叶癫痫定侧定位诊断上的应用

Statistical parametric mapping-based interhemispheric asymmetry analysis of ^18F-FDG PET brain imaging in localization of unilateral temporal lobe epileptic foci

目的评价基于SPM的脑内不对称分析法在^18F－FDG PET代谢显像对单侧难治性颞叶癫（TLE）患者致灶辅助定侧定位诊断中的应用价值。方法回顾性分析2012年3月至2015年7月间73例[男42例、女31例，平均年龄（27.10±9.78）岁]单侧TLE患者及30名[男17名、女13名，平均年龄（33.79±5.46）岁]健康对照组的临床及影像学资料。通过^18F－FDG PET/CT检查对TLE患者主要致灶进行定侧定位，^18F－FDG PET图像分析采用基于SPM的脑内不对称分析法、视觉分析法及不对称指数（AI）分析法。综合术后病理及预后，比较3种方法对单侧TLE患者主要致灶的检出符合率。数据分析采用χ^2分析及χ^2分割检验。结果运用基于SPM的脑内不对称分析法、视觉分析法及AI分析法对单侧TLE患者定侧定位诊断符合率分别为87.67%（64/73）、75.34%（55/73）及69.86%（51/73），差异有统计学意义（χ^2＝6.993，P〈0.05）；两两比较的结果显示，脑内不对称分析法与AI分析法、脑内不对称分析法与视觉分析法间差异有统计学意义（χ^2值：8.481和5.818，均P〈0.017），视觉分析法与AI分析法间差异无统计学意义（χ^2＝0.500，P〉0.017），脑内不对称分析法的符合率明显高于视觉分析法及AI分析法。结论基于SPM的脑内不对称分析法可以相对客观、准确地对单侧难治性TLE患者的主要致灶进行定位，有较高的临床应用价值。

Objective To evaluate the diagnostic value of the SPM-based interhemispheric asymmetry analysis （SIAA） of ^18F-FDG PET brain imaging in localization of temporal lobe epileptic foci. Methods A total of 73 patients （42 males, 31 females, average age： （27.10±9.78） years） diagnosed as temporal lobe epilepsy （TLE） and 30 healthy volunteers （17 males, 13 females, average age： （33.79±5.46） years） from March 2012 to July 2015 were enrolled in this retrospective study. ^18F-FDG PET brain imaging was performed on all subjects to localize temporal lobe epileptic foci. The results of ^18F-FDG-PET were analyzed by SIAA, visual analysis and asymmetry index （AI） analysis. The accuracy of the three methods was evaluated according to the postoperative pathological results and curative effect. Х^2 test and partition of Х^2 test were used to analyze the data. Results The accuracies of localization with SIAA, visual analysis and AI analysis were 87.67% （64/73） , 75.34% （ 55/73 ） and 69.86% （ 51/73 ） , respectively （Х^2 = 6.993, P〈 0. 05 ）. The accuracy of SIAA was higher than that of visual analysis and AI analysis （Х^2 values： 5.818 and 8.481, both P〈0.017）, while there was no significant difference between AI analysis and visual analysis （Х^2=0.500, P〉0.017）. Conclusion The SIAA of ISF-FDG PET brain imaging may be a more objective and accurate method in localization of TLE foci.