PURPOSE: To assess two-dimensional phase-encoded proton chemical shift imaging for potential clinical application in presurgical localization of temporal lobe epilepsy (TLE). MATERIALS AND METHODS: Interictal chemical shift imaging studies were performed in 25 patients (17 unilateral, eight bilateral epileptogenic abnormalities) and 12 healthy volunteers. Results were compared with those of electroencephalography. For each temporal lobe, the volume of interest (VOI) included mesial, lateral, anterior, and posterior areas. RESULTS: The epileptogenic zone was characterized by an abnormal pattern of decreased signal intensity in the N-acetylaspartate (NAA) peak, either increased or unchanged choline (Cho) level relative to creatine (Cr) signal intensity, and, occasionally, elevated lactate level. NAA/Cho was the most sensitive and reliable quantitative marker for abnormality. The sensitivity and specificity of this technique compared with EEG were 90% and 85%, respectively. The difference in NAA/Cho between epileptogenic and normal temporal lobes was highly significant (P < .001). No statistically significant difference was found between normal temporal lobes and patients' uninvolved temporal lobes. CONCLUSION: NAA/Cho is an excellent marker for localizing the epileptogenic zone in TLE.
PURPOSE: To assess two-dimensional phase-encoded proton chemical shift imaging for potential clinical application in presurgical localization of temporal lobe epilepsy (TLE). MATERIALS AND METHODS: Interictal chemical shift imaging studies were performed in 25 patients (17 unilateral, eight bilateral epileptogenic abnormalities) and 12 healthy volunteers. Results were compared with those of electroencephalography. For each temporal lobe, the volume of interest (VOI) included mesial, lateral, anterior, and posterior areas. RESULTS: The epileptogenic zone was characterized by an abnormal pattern of decreased signal intensity in the N-acetylaspartate (NAA) peak, either increased or unchanged choline (Cho) level relative to creatine (Cr) signal intensity, and, occasionally, elevated lactate level. NAA/Cho was the most sensitive and reliable quantitative marker for abnormality. The sensitivity and specificity of this technique compared with EEG were 90% and 85%, respectively. The difference in NAA/Cho between epileptogenic and normal temporal lobes was highly significant (P < .001). No statistically significant difference was found between normal temporal lobes and patients' uninvolved temporal lobes. CONCLUSION:NAA/Cho is an excellent marker for localizing the epileptogenic zone in TLE.