Abstracts

Rationale: In temporal lobe epilepsy (TLE), a common type of epilepsy in adults, hippocampal changes are frequently observed in magnetic resonance imaging (MRI), such as reduced volume and increased T2-weighted signal. Ex vivo MRI studies in brain specimens have shown that the same signal alterations observed in vivo can be seen in fixed tissue. Our aim was to evaluate the hippocampus of TLE patients in vivo and ex vivo conditions using two quantitative magnetic resonance techniques (T2 relaxometry and Magnetization Transfer) Methods: Patients with TLE (n = 18) were scanned prior to surgical resection of the hippocampus (in vivo imaging) in a 3 T MRI scanner. For T2 relaxometry, spin echo images with multiples echoes (TE = 20, 40, 60, 80 and 100 ms; TR = 3 s) were acquired and post-processed. Other two images, gradient echo (TR = 7.3 ms, TE = 3.6 ms, with/without magnetization transfer pulse), were used to estimated the Magnetization Transfer Ratio (MTR) maps. After surgery, a 2 mm thick coronal section from hippocampal body was fixed for 8 days, and very similar ex vivo images were performed in the same scanner. Age-matched controls were obtained from voluntaries (radiological control, RC, n = 12) for the in vivo measurements, and from autopsy tissues (histological and radiological control, HRC, n = 6) for the ex vivo measurements. Results were considered significant at p < 0.05 Results: TLE patients presented increased T2 values (TLE = 117 ms, RC = 104 ms; median, p = 0.005) and reduced MTR (TLE = 48 ± 2 %, RC = 50 ± 1%; mean ± standard error, p = 0.012) when compared to RC. In qualitative terms, ex vivo imaging had a better anatomical definition than in vivo imaging. This better anatomical definition allowed us to identify and quantify the T2 values in hippocampal neuronal layers (NL), radiatum/oriens layers (ROL) and molecular layers of fascia dentata (ML) with great confidence. Ex vivo relaxometry revealed increased T2 in TLE, compared to HRC, in the whole hippocampus (TLE = 83 ms, HRC = 76 ms; median, p = 0.006) and in the subdivisions NL (TLE = 85 ms, HRC = 78 ms; median, p = 0.006), ROL (TLE 80 ms, HRC = 72 ms; median, p = 0.033) and ML (TLE = 84 ms, HRC = 78 ms; median, p = 0.008). No correlation was observed between in vivo and ex vivo conditions for T2 values in the whole hippocampus neither in the several regions Conclusions: Ex vivo acquisitions provided an excellent anatomical imaging of the hippocampus. In vivo and ex vivo T2 values were increased in TLE patients, but no correlation was observed between both conditions. While the alterations in the in vivo T2 are related to gliosis and tissue edema, the ex vivo alteration must be purely structural, since no edema remains after the fixation process. These differences between the substrates related to in vivo and ex vivo signals could be responsive to the lack of correlation in the relaxometry technique. Taken together, our data suggest that structural and functional changes account for MR signal, and ex vivo MRI can identify some of these alterations