Open Access Research article

'MRI-negative PET-positive' temporal lobe epilepsy (TLE) and mesial TLE differ with quantitative MRI and PET: a case control study

Ross P Carne1103479*, Terence J O'Brien1046, Christine J Kilpatrick1046, Lachlan R MacGregor48, Lucas Litewka1, Rodney J Hicks102 and Mark J Cook11035

Author Affiliations

1 Victorian Epilepsy Centre, St Vincent's Hospital, Melbourne, Australia

2 PET Centre, The Peter MacCallum Cancer Institute, Melbourne, Australia

3 Department of Neurology, St. Vincent's Hospital, Melbourne, Australia

4 Department of Neurology, The Royal Melbourne Hospital, Melbourne, Australia

5 Department of Medicine, St. Vincent's Hospital, Melbourne, Australia

6 Department of Medicine, The Royal Melbourne Hospital, Melbourne, Australia

7 Department of Neuroscience, The Geelong Hospital, Barwon Health, Geelong, Australia

8 Department of Clinical Epidemiology and Biostatistics, The Royal Melbourne Hospital, Melbourne, Australia

9 School of Medicine, Deakin University, Geelong, Australia

10 Faculty of Medicine, The University of Melbourne, Melbourne, Australia

For all author emails, please log on.

BMC Neurology 2007, 7:16  doi:10.1186/1471-2377-7-16

Published: 24 June 2007



'MRI negative PET positive temporal lobe epilepsy' represents a substantial minority of temporal lobe epilepsy (TLE). Clinicopathological and qualitative imaging differences from mesial temporal lobe epilepsy are reported. We aimed to compare TLE with hippocampal sclerosis (HS+ve) and non lesional TLE without HS (HS-ve) on MRI, with respect to quantitative FDG-PET and MRI measures.


30 consecutive HS-ve patients with well-lateralised EEG were compared with 30 age- and sex-matched HS+ve patients with well-lateralised EEG. Cerebral, cortical lobar and hippocampal volumetric and co-registered FDG-PET metabolic analyses were performed.


There was no difference in whole brain, cerebral or cerebral cortical volumes. Both groups showed marginally smaller cerebral volumes ipsilateral to epileptogenic side (HS-ve 0.99, p = 0.02, HS+ve 0.98, p < 0.001). In HS+ve, the ratio of epileptogenic cerebrum to whole brain volume was less (p = 0.02); the ratio of epileptogenic cerebral cortex to whole brain in the HS+ve group approached significance (p = 0.06). Relative volume deficits were seen in HS+ve in insular and temporal lobes. Both groups showed marked ipsilateral hypometabolism (p < 0.001), most marked in temporal cortex. Mean hypointensity was more marked in epileptogenic-to-contralateral hippocampus in HS+ve (ratio: 0.86 vs 0.95, p < 0.001). The mean FDG-PET ratio of ipsilateral to contralateral cerebral cortex however was low in both groups (ratio: HS-ve 0.97, p < 0.0001; HS+ve 0.98, p = 0.003), and more marked in HS-ve across all lobes except insula.


Overall, HS+ve patients showed more hippocampal, but also marginally more ipsilateral cerebral and cerebrocortical atrophy, greater ipsilateral hippocampal hypometabolism but similar ipsilateral cerebral cortical hypometabolism, confirming structural and functional differences between these groups.