Abstracts

Rationale: A considerable number of Temporal Lobe Epilepsy (TLE) patients were found to have normal high-resolution MRI of the hippocampi despite EEG findings and clinical evaluation consistent with TLE. This prompted us to add In vivo proton MRS and volumetric measurements of the hippocampi.Methods: One hundred and thirty four patients (61 female, 73 males, ages 19-43, mean age 29 years) and 26 age-matched healthy volunteers were examined with high resolution MRI and In vivo Proton MRS of the hippocampi. A 1.5 Tesla MRI unit (Signa HDx, General Electric, Milwaukee, USA) was used. Sagittal T1-, axial FLAIR and diffusion-weighted images (DWI), and high resolution coronal T2-w, T1-FLAIR and T2-FLAIR images were acquired. Manual shimming was virtually always used for improving automatic shimming. A quadrature head coil was used. Two spectra at each hippocampus were acquired with TR/TE = 1500/35 and 135 msec. The absolute concentrations of N-Acetyl Aspartate, choline (Cho), creatine and phosphocreatine (Cr) and Myo-Inositol (MI) were determined with a method based on an external phantom of known metabolite concentrations. Volumetric measurements were performed on Sagittal T1-w and coronal T1-FLAIR images of 29 patients. Results: Compared to normal volunteers 94% of the patients had statistically significant reduction of the NAA/Cr ratio in both hippocampi and only 6% of the patients had one normal hippocampus. Comparing left to right hippocampus in the same patient, 54% of the patients had statistically significant differences between left and right hippocampus and 46% of the patients had bilateral pathology of equal severity. Similar findings were found for Cho/Cr ratios. The MI/Cr ratio was not always in agreement with the findings of NAA/Cr and Cho/Cr. The volumetric measurements were in good agreement with the MRS results in 81% of the cases compared. In the remaining 19% of the cases either the hippocampal volumes were not reduced in a statistically significant way compared to normal volunteers or lateralization could not be accomplished. Confirming the MRS results with the volumetric measurements makes a stronger case for the determination of the laterization of the disease and helps in the selection of those patients who might benefit the most from a potential surgery. Our results help explain the current rate of success (roughly 50%) for seizure-free post-surgery patients. With a better evaluation of candidates for surgery the rate of success might increase. The combination of In Vivo proton MRS and Volumetric measurements of hippocampi seems to be a potent probe for TLE. Plain MRI, including state-of-the-art high resolution techniques, used in a qualitative way is not sufficient in all cases. Conclusions: The most important finding of our study was the bilateral pathology found in 94% of all TLE patients. Having one normal hippocampus in TLE is rather the exception (only 6% of the patients). These results add significant information for the pre-surgical evaluation of patients.