Diagnosis of medial temporal lobe seizure onset: relative specificity and sensitivity of quantitative MRI

Neurology. 1993 Oct;43(10):2117-24. doi: 10.1212/wnl.43.10.2117.

Abstract

Measurement of hippocampal volume by MRI is a new technique with potential value in the localization of epileptic regions, but whether reduced hippocampal volume predicts the location of electrical seizure onset in mixed patient groups is not known. We examined the sensitivity and specificity of this technique among 56 refractory epileptic patients for the diagnosis of medial temporal lobe epilepsy as judged by intracranial EEG recording of spontaneous seizure onset. Since these patients had intracranial EEG because of inconsistent or insufficient localization by noninvasive electrophysiologic, functional, and structural assessment, this patient population can be considered the most difficult to localize. Hippocampal atrophy by MRI volumetry was 75% sensitive to, and 64% specific for, ipsilateral medial temporal lobe seizure onset in this group. Hippocampal atrophy was significantly correlated with longer duration of epilepsy. MRI compared favorably with all other noninvasive means of localization, which had 41 to 73% sensitivity and 45 to 65% specificity to medial temporal ictal onset; of these, none had as acceptable a combination of both adequate sensitivity and specificity. We conclude that MRI volumetry of the hippocampus is a valuable, noninvasive localization method in chronic epilepsy, with a yield and accuracy surpassing all other noninvasive studies used to predict the presence of medial temporal seizure onset.

Publication types

  • Comparative Study
  • Research Support, U.S. Gov't, Non-P.H.S.

MeSH terms

  • Atrophy
  • Electroencephalography
  • Epilepsy, Temporal Lobe / diagnosis*
  • Epilepsy, Temporal Lobe / pathology
  • Epilepsy, Temporal Lobe / physiopathology
  • Fever
  • Hippocampus / pathology*
  • Hippocampus / physiopathology
  • Humans
  • Magnetic Resonance Imaging / methods*