Syndromic Diagnosis at the Epilepsy Clinic: Role of MRI in Lobar Epilepsies

Abstract

PURPOSE: Magnetic resonance imaging (MRI) is an essential diagnostic tool for the management of epilepsy at modern epilepsy clinics. This study was conducted to incorporate MRI features into the international classification of epilepsies and epilepsy syndromes (ICEES) proposed by the International League Against Epilepsy (ILAE). METHODS: Three hundred consecutive patients newly registered in the Yonsei Epilepsy Clinic underwent stepwise classifications based on clinical features, clinical EEG, and clinical EEG-MRI correlations. The patients were required to have epilepsy and have undergone both EEG and MRI for inclusion in the study. Interictal epileptiform discharges (IEDs) in the EEG were divided into lobar, multilobar, and generalized. MRI lesions were divided into lobar and multilobar lesions. Lobar epilepsies (LEs) were divided into temporal, frontal, parietal, occipital, rolandic, temporoparietooccipital junctional, multilobar, and nonlocalized LEs. RESULTS: Two hundred forty-nine patients satisfied the inclusion criteria. In the first-step diagnosis, 190 patients were classified as having localization-related epilepsy; 24 patients, generalized epilepsy; 34 patients, undetermined epilepsy; and one patient, a special syndrome. EEG revealed IEDs in 124 (50%) patients, and the second-step diagnosis changed the diagnostic categories of 79 (32%) patients. MRI detected lesions in 106 (43%) patients, and the third-step diagnosis changed the diagnostic categories of 30 (12%) patients. The nonspecific diagnostic categories of ICEES decreased from 49% to 37% and then to 29%, as diagnosis progressed from steps one to three. In cases of LE, MRI was superior to EEG in its clinical correlation. Additionally, the diagnostic precision in temporal lobe epilepsy was far better than that for other LEs. CONCLUSIONS: The impact of MRI on ICEES was only modest in terms of changing diagnostic categories, although MRI provided a structural substrate for epilepsy in 38% of patients with negative EEGs. In LE, MRI was as sensitive as EEG, and its clinical correlation was superior to that of EEG, which strongly supports the rationale of incorporating MRI into ICEES.