102 Archives ofDisease in Childhood 1996;75:102-107 ORIGINAL ARTICLES Arch Dis Child: first published as 10.1136/adc.75.2.102 on 1 August 1996. Downloaded from Focal abnormalities detected by '8FDG PET in epileptic encephalopathies Colin D Ferrie, Michael Maisey, Timothy Cox, Charles Polkey, Sally F Barrington, Chrystosomos P Panayiotopoulos, Richard 0 Robinson Abstract cal treatment not only for control of seizures A prospective study of 32 children with but possibly also to improve neurodevelop- epileptic encephalopathies 12 years or mental outcome.4 Seizures in the epileptic younger revealed a high incidence of focal encephalopathies are usually considered to be cortical metabolic defects on 18- primary generalised, arising from a cortex with fluorodeoxyglucose positron emission to- diffuse or multifocal abnormalities. Despite mography (PET) not suspected from this, cases are described in patients with readily clinical, EEG, or magnetic resonance detectable localised brain lesions including imaging findings. PET scans were normal tumours, porencephalic cysts, and severe mi- in all five children with typical de novo grational anomalies such as pachygyrias.' Lennox-Gastaut syndrome but showed Seizures are associated with pronounced cortical metabolic abnormalities in three changes in the metabolism of substrates such out of four with atypical de novo Lennox- as glucose. In partial epilepsies focal reduction Gastaut syndrome, five out of six with (hypometabolism) and increase in glucose Lennox-Gastaut syndrome following in- consumption (hypermetabolism) in the epilep- fantile spasms, six out ofeight with severe togenic zone is characteristic of the interictal myoclonic epilepsy in infancy, one out of and ictal states respectively.6 This can be two with epilepsy with myoclonic-astatic detected by positron emission tomography seizures, and four out of six with an (PET) using 18-fluorodeoxyglucose (FDG), a unclassified epileptic encephalopathy. radionucleotide labelled analogue of glucose. http://adc.bmj.com/ This suggests that some children with epi- Recently FDG PET studies have contributed leptic encephalopathies previously to the detection of subtle localised abnormali- thought to have primary generalised sei- ties of neuronal migration in children with cer- zures or seizures due to multifocal pathol- tain epileptic encephalopathies with apparently ogy may have unifocal cortical origin for generalised seizures.7-" In patients with infan- Department of their seizures. Such an origin may be tile spasms this has lead to cortical resections Paediatric Neurology, amenable to surgery. with excellent seizure control and improved on September 26, 2021 by guest. Protected copyright. Guy's Hospital, London (Arch Dis Child 1996;75:102-107) development."214 C D Ferrie However, these studies used broad defini- R 0 Robinson Keywords: epilepsy, positron emission tomography, tions, particularly of the Lennox-Gastaut infantile spasms, Lennox-Gastaut syndrome. and often included with Clinical PET Centre, syndrome, patients UMDS, London clinical, EEG, or structural neuroimaging M Maisey evidence strongly indicative of focal brain S F Barrington The term 'epileptic encephalopathy' encom- passes various severe childhood epilepsies pathology, making the contribution of PET Department of characterised by multiple seizure types and dif- unclear. In earlier studies computer assisted Neurological Sciences, fusely slow EEG with generalised or multifocal tomography rather than magnetic resonance Guy's and St Thomas' imaging (MRI) scans was often used, and even Hospitals, London paroxysmal abnormalities. Psychomotor delay T Cox with onset preceding or following the onset of when MRI was employed, it is not clear if the seizures is usual.' 2 The best known of these technical specifications were sufficient to allow Maudsley Hospital, conditions are West's and the Lennox-Gastaut detection of subtle abnormalities. In this study London we aimed to establish whether FDG PET C Polkey syndromes (International League Against Epi- lepsy code 2.2).? Others include epilepsy with detects focal abnormalities in children with Department of Clinical myoclonic-astatic seizures (2.2) and severe epileptic encephalopathies in whom this is not Neurophysiology and clear from clinical examination, EEG, or high Epilepsy, St Thomas' myoclonic epilepsy in infancy (3.2).' Their Hospital, London intractable nature, associated developmental resolution MRI studies, and whether there are C P Panayiotopoulos problems, and frequent need for residential syndrome related differences in the PET find- schooling makes them important conditions ings. Correspondence to: Dr C D Ferrie, Newcomen both to paediatric neurologists and to general Centre, Guy's Hospital, and London SE1 9RT. community paediatricians. Current medi- Methods cal treatment is usually disappointing and has Children of 1 to 12 years of age with an epilep- Accepted 17 May 1996 prompted renewed interest in the role of surgi- tic encephalopathy of unknown aetiology were "8FDG PET in epileptic encephalopathies 103 studied prospectively. Treatment with at least discouraged. Intravenous diazepam was given three major anticonvulsants had failed in all to patients with frequent paroxysmal EEG and they were considered potential candidates activity before or during the first 10 minutes of Arch Dis Child: first published as 10.1136/adc.75.2.102 on 1 August 1996. Downloaded from for surgery. Epileptic encephalopathy was FDG uptake, and also to agitated patients. defined as the occurrence of mixed generalised Patients were scanned after 30 minutes of seizures (tonic, atonic, myoclonic, atypical FDG uptake, using a head holder to minimise absence, generalised tonic-clonic) with, in the movement. Six 5 minute consecutive frames fully evolved syndrome, a diffusely slow EEG were acquired with the data summed. Frames with generalised or multifocal interictal parox- with excessive movement were discarded. Fol- ysmal abnormalities and (if recorded) general- lowing correction for attenuation, images were ised ictal abnormalities. Patients with seizures smoothed and reconstructed to give 31, 3.4 which were difficult to classify but which did mm thick planes with an in plane spatial reso- not have localising or lateralising value and lution of 8 mm and a total axial field ofview of those with ictal symptoms of uncertain localis- 10.4 cm. The images were reconstructed in ing value-such as eye deviation during gener- axial and coronal planes and in the plane alised seizures-were included provided other parallel to the long axis of the temporal lobes. typical seizures were also present. Patients with MRI and PET images were inspected 'soft' or equivocal focal neurological signs were visually blind to the clinical information and to eligible. each other. Semiquantitative analysis of PET Patients with seizures with localising value data was performed using a template of multi- were excluded except if they occurred before ple 4 mm diameter circular regions of interest the onset of infantile spasms or in those classi- (ROI) placed in selected areas chosen by fied as severe myoclonic epilepsy in infancy in matching PET planes to an anatomical brain whom focal ictal symptoms are characteristic.3 atlas. ROI were placed on anatomical grounds Patients whose EEG showed consistent asym- rather than on functional imaging findings; metries or focal abnormalities were excluded, thus areas of apparent increased or decreased as were those with clear focal neurological FDG uptake were not preferentially selected. deficits and those in whom previous neuroim- Frontal, parietal, occipital, medial, and lateral aging showed focal abnormalities or other temporal and cerebellar cortices as well as len- abnormalities likely to preclude surgery. Spe- tiform and caudate nuclei, thalami, mid-brain, cific metabolic conditions were excluded by and pons were sampled. Control PET data for measurement of plasma biotinidase, lactate, normal children is lacking for ethical reasons. pyruvate, ammonia and amino acids, urine FDG PET studies in adults with partial organic acids, cerebrospinal fluid (CSF) lac- seizures suggest an asymmetry of 15% or more tate, pyruvate, and glycine, the CSF:blood glu- in homologous cortical regions is abnormal.23 cose ratio (for glucose carrier protein defi- Studies in children show important regional ciency),"6 and by a trial of oral pyridoxine. changes during development but did not A detailed clinical review, examination find- suggest greater side to side asymmetries.24 We http://adc.bmj.com/ ings, and EEG including sleep and video used an asymmetry of greater than 15% to recordings was used to classify patients syndro- define unequivocal abnormality. Detection of mically according to recommendations of the bilateral and diffuse abnormalities on semi- International League Against Epilepsy supple- quantitative analysis is difficult. We report it mented by those of leading authorities.3 17-22 only when obvious on visual inspection. Close Additionally, patients were divided according observation of the patient and the EEG to whether they were typical or atypical of the recorded during radiotracer uptake was used to particular syndrome. Features leading to a des- classify scans as ictal (paroxysmal EEG activity on September 26, 2021 by guest. Protected copyright. ignation as atypical included onset of seizures generally acknowledged as ictal with or without outside the age limits considered normal for clinical events) or as interictal (no ictal-type