MRI OF THE FORNIX AND MAMILLARY BODY IN TEMPORAL-LOBE EPILEPSY

We performed MRI on 27 patients with clinically proven temporal lobe e pilepsy (TLE), all with prior EEG lateralisation, and 10 volunteers, s tudied to evaluate disparity in size arising from biological variation (group 1). Three-dimensional spoiled GRASS (3DSPGR) sequences provide d 2-mm contiguous sections of the limbic system, enabling assessment o f the hippocampus (HC), fornix (FN) and mamillary body (MB). Measureme nts of FN and MB width were made from a workstation. Any percentage di fference in size was computed. In 19 cases there was unilateral abnorm ality in the HC (group 2); in 18 and 19 cases respectively there was a smaller FN and MB on the same side as the abnormal HC. This percentag e difference in size was significantly greater than that in group 1 in the FN and MB in 17 and 17 cases respectively. Comparison of percenta ge difference computations for FN and MB between groups 1 and 2 showed high statistical significance (P < 0.0002). In 5 patients with clinic al TLE the HC was normal on MRI (group 3). Unequal FN and MB sizes wer e found in 4, significant in 2. Comparison of percentage difference co mputations for FN and MB showed statistical significance (P < 0.0005 a nd P < 0.0003 respectively). There was no case of discordance between the sides of hippocampal abnormality and the smaller FN or MB or betwe en the sides of smaller FN and MB. The strong concordance between the changes in the HC and those in the FN and MB suggests that this combin ation will play an important role in the assessment of TLE and limbic system abnormality.