^(1)H-MRS在颞叶癫癎定侧诊断中的价值

valuation of Proton Magnetic Resonance Spectroscopy for Lateralization of Temporal Lobe Epilepsy

目的：探讨１Ｈ－ＭＲＳ在颞叶癫 （ＴＬＥ）定侧诊断（ＬＤ）中的价值。材料和方法：２０例手术和病理证实的顽固性颞叶癫 患者和１５例健康志愿者为研究对象。患者术前采用 １～７天的视频脑电连续监护或（和）颅内电极脑电图（ＥＥＧ）行定侧诊断。采用ＧＥ１．５Ｔ Ｓｉｇｎｓ Ｈｏｒｉｚｏｎ ＬＸ超导磁共振机，作双侧颞叶内侧的１Ｈ－ＭＲＳ，感兴趣区大小为２ｃｍ ×２ｃｍ × ２ｃｍ。将正常对照组ＮＡＡ／（Ｃｒ＋Ｃｈｏ）比值的９５％可信区间的下限值和双侧ＮＡＡ／（Ｃｒ＋Ｃｈｏ）比值差别的最大值定为１Ｈ－ＭＲＳ ＴＬＥ ＬＤ的标准。分析１Ｈ－ＭＲＳ作为ＴＬＥ ＬＤ的价值，并与ＥＥＧ、病理、 ＰＥＴ及ＭＲＩ所见相对照。结果：正常对照组ＮＡＡ／（Ｃｒ＋Ｃｈｏ）比值平均为０．８１±０．１３（０．７０～１．０５）；双侧ＮＡＡ／（Ｃｒ＋Ｃｈｏ）比值的差别平均为５％±２％（１％～１０％）。以ＮＡＡ／（Ｃｒ＋Ｃｈｏ）＜０．６８和双侧ＮＡＡ／（Ｃｒ＋Ｃｈｏ）比值的差别＞７％作为ＴＬＥ ＬＤ的标准。手术切除的颞叶均有病理改变，术后随访２０例均有疗效，表明ＥＥＧ定侧全部正确。２０例中的１６例（８０％）１Ｈ－ＭＲＳＬＤ与ＥＥＧ相符，１例不符，另外３例１Ｈ－

To evaluate the value of proton magnetic resonance spectroscopy (1H-MRS) in the lateralization diagnosis (LD) of temporal lobe epilepsy (TLE).Materials and Methods:20 patients with intractable TLE verified by operation and pathology and 15 healthy volunteers were studied. Before partial temporal lobectomy, continuous video EEG monitoring for 1 - 7 days and/or intracranial EEG were used for the LD of TLE. Using GE 1.5T Signa Horizon LX MRI machine, spectra were obtained from 2cm × 2cm × 2cm cubes in the medial region of both temporal lobes. The lower threshold of the 95%confidence level of the NAA/(Cr+Cho) ratio and the highest NAA/(Cr+Cho)difference between the two sides obtained from the normal control group were arbitrarily specified as the standards for the LD of TLE. The diagnostic value of 1H-MRS for the LD of TLE was analyzed and compare4 with which of MRI,PET and EEG. Results: The NAA/(Cr+Cho) ratio of the normal control group was 0.81 ± 0. 13(0.70 - 1 .05),and the difference of NAA/(Cr+Cho) ratio of two sides was 5% ± 2%(1% - 10%). We arbitrarily adopted the NAA/(Cr+Cho)<0.68 and >7% of the difference of NAA/(Cr+Cho) of two sides as the standards for the LD of TLE. All excised temporal lobes showed pathological changes. The surgery was proved to be effective by follow -ups of all the 20 patients. Both results demonstrated that the LD of all EEG were correct. The IH-MRS LDs were same as which of EEG in 16 cases (80%)and different in lease (5%)respectively. 1H-MRS was negative in the other 3cases(15%).MRI abnormalities were found in 18cases. MRIs were negative in other 2cases, but their 1H-MRS were positive. PET of 14.cases appeared as decreased uptake of FDG,13were found in single side and 1 was found in both sides.Among these 14 cases,LD of PET and 1H-MRS were different with which of EEG in 1 and 2 cases respectively. Both sides abnormalities were found more often in 1H-MRS(5 cases) than which in PET. We also found that patients with bilateral MRS abnormalities had a longer history of seizure than those with unilateral abnormal and normal group (p<0.05). Conclusion:1H-MRS in the LD of TLE is rather accurate.When 1H-MRS combined with MRI, the accuracy of LD can be elevated further. Since both video EEG and PET are expensive and time consuming, 1H-MRS in conjunction with MRI can partially replace EEG and PET.