目的：探讨18F-FDG PET与MRI脑代谢显像对颞叶癫痫（TLE）致痫灶定侧的诊断价值。方法：对78例资料完整的难治性TLE患者进行回顾性分析，以病理结果为金标准，对其常规MRI、MRS、18F-FDG PET检查及三者联合诊断在TLE定侧上的检出准确率进行比较、分析，以评价其价值。结果：本组78例难治性TLE患者均行前述三种检查，经统计学分析，致痫灶侧NAA/（Cho+Cr）及SUVmean低于健侧，三种检查定侧检出准确率分别为33.33%（26/78）、58.97%（46/78）及85.89%（67/78），三者联合诊断准确率为89.74%（70/78），经卡方检验分析，得到χ2=73.090，P<0.001，可以认为总体差异有统计学意义，采用分割卡方进一步的两两比较，对P值进行校正，以0.008为标准，可以认为常规MRI与MRS（χ2=10.317，P=0.001）、18F-FDG PET（χ2=44.758，P<0.001）、三者联合（χ2=52.433，P<0.001）差异均有统计学意义；MRS与18F-FDG PET（χ2=14.158，P<0.001）、三者联合（χ2=19.366，P<0.001）差异均有统计学意义；尚不能认为18F-FDG PET与三者联合（χ2=0.539，P=0.463）差异有统计学意义。结论：MRS定侧检出准确率明显高于常规MRI，在临床癫痫影像诊断时可先利用MRI进行诊断，但应尽量同时行常规MRI及MRS检查，若经前述检查仍无法准确定位，有条件者可进行18F-FDG PET检查，三者联合应用可显著提高癫痫致痫灶定侧检出准确率。

Objective: To evaluate the value of 18F-FDG PET and MRI in the localization of temporal lobe epileptic focus. Methods: Seventy-eight patients diagnosed as TLE underwent conventional MRI, MRS and 18F-FDG PET brain imaging. With the pathological results as the gold standard, the accuracies of these three methods and combined application of these three methods were analyzed to evaluate the value of the methods mentioned above. Results: After statistical analysis, the values of NAA/(Cho+Cr) and SUVmean in the affected side were lower than that of the healthy side. The accuracy rates of these three kinds examinations and combined application of these three methods in localization of temporal lobe epileptic focus were 33.33%(26/78), 58.97%(46/78), 85.89%(67/78)  and 89.74%(70/78) respectively. After the chi-square test, there was statistically significant difference(χ2=73.090, P<0.001). Using the chi-square decomposition to further pairwise comparison with the P value corrected to 0.008 as the standard, the differences between conventional MRI and MRS, 18F-FDG PET imaging, combined application were statistically significant(χ2 was 10.317, 44.758, 52.433, respectively and P=0.001, P<0.001, P<0.001, respectively).The differences between MRS and 18F-FDG PET, combined application were statistically significant(χ2 was 14.158, 19.366, respectively and P values are both less than 0.01). The difference between the 18F-FDG PET and the combined application was statistically significant(χ2=0.539, P=0.463). Conclusion: MRI can be used in the clinical diagnosis of epilepsy, but the conventional MRI and MRS examination should be carried out at the same time, because the accuracy of MRS was significantly higher than that of conventional MRI. Patients should be examined by 18F-FDG PET imaging if the patients were still unable to be diagnosed by the above examination and were willing to accept the examination. The combination of the three methods in localization of temporal lobe epileptic focus can significantly improve the diagnostic accuracy of intractable temporal lobe epilepsy.