REVISTA DO DETUA, VOL. 3, Nº 7, SETEMBRO 2002 Abstract – Chemical-shift imaging (CSI) has been applied to the evaluation of patients with temporal lobe epilepsy (TLE), but few comparisons have been made between the data obtained from quantitative structural magnetic resonance imaging (MRI) and the accuracy of CSI to identify pathological changes within the hippocampus. Our goal was to determine the discriminative capability of CSI hippocampal NAA/(Cho+Cre) ratios to distinguish between patients with lateralized epileptic foci when compared to structural MRI measures of disease. I. INTRODUCTION Magnetic resonance has assumed an increasingly important role in the study of patients with epilepsy that now extends beyond its demonstration of neuroanatomy and neuropathology into the fields of neurochemistry, neurometabolism and neurophysiology. The use of chemical-shift imaging (also know as Proton multi-voxel spectroscopy – 1 H MRS) to obtain in vivo biochemical information from the brain of living, awake subjects provides a powerful tool in the investigation of tissue metabolism and complements the anatomical information delivered by MR imaging. Specifically, the pathological hallmark of the sclerotic and dysfunctional hippocampus in TLE has been analysed by the relative ratios of metabolite resonance signals such as total N- acetylaspartate (NAA), total creatine plus phosphocreatine (Cr) and choline-containing compounds (Cho). In patients with TLE and hippocampal sclerosis (HS), CSI has frequently shown abnormalities which consist mostly in the reduction of NAA. It is believed that NAA represents a robust but unspecific marker for neuronal loss or dysfunction. According to the former hypothesis, a decrease in NAA would be directly correlated with the degree of cell loss as seen in HS, while in the latter its value would also be altered by transient metabolic states experienced by an active neuronal network, either in a disease or healthy state. The good concordance between the side of maximal NAA reduction and the side of seizure origin is generally accepted [1, 2]. Still, the metabolic abnormality can be more extensive than the epileptogenic area defined by electroencephalography (EEG) or structural imaging [3]. It has been shown that 18% [1] -50% [3] of patients with unilateral TLE, can present a bilateral reduction in the relative metabolic ratios (either NAA/Cr or NAA/Cho+Cre) which agrees with some reported post- mortem cases [4] where asymmetric cell loss was quantitatively verified in both hippocampi, with a higher degree ipsilateraly but affecting also the opposite hippocampus . Thus, caution is necessary while interpreting the spectroscopic profile obtained from TLE as bilateral abnormalities can obscur the relevant epileptogenic and surgically ressecable lesion. Volumetric measures of the hippocampal formation (Hvol) has provided a well establish technique to assess focal atrophies within any part of the hippocampus [5, 6]. A high degree of pathological specificity is associated with the finding of even minor volume asymmetries [7, 8] and a good correlation between hippocampal atrophies and surgical outcome has been reported [9]. Comparatively with CSI-data, MRI volume measurements had highlighted few bilateral damaged hippocampi (9- 18%) in the majority of reported surgical series [10, 11], which happens to have no clear relation with the clinical outcome [11]. Still, verbal memory outcome was seriously prejudiced if bilateral atrophy was associated with left focus TLE [12]. The principal MRI features of HS are volume loss and increased T2 weighted signal intensity. In general, the increased signal is thought to be related to reactive gliosis occurring within a damaged brain area. In the epileptogenic hippocampus, increased T2-relaxometry (HT2) has been specifically associated with gliosis in the dentate gyrus [13]. Since abnormal T2 signal intensity may be difficult to detect visually, HT2 has been validated as an objective Comparative study of chronic TLE patients using structural and metabolic MRI measures of hippocampal damage. P.M. Goncalves Pereira 1 , M. Forjaz Secca 2,3 , A.Leal 4 , P. Evangelista 1,2 , C. Ribeiro 1,2 , P. Rosado 5 1. Serviço de Neurorradiologia, Hospital de Egas Moniz, Lisboa. 2. Ressonância Magnética de Caselas, S.D.I. Lisboa. 3. Dpt. de Física, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa. 4. Serviço de Neurologia, Hospital de Fernando Fonseca, Amadora. 5. Serviço de Neurologia, Hospital de Egas Moniz, Lisboa.