Abstracts

MRI is better than CT at identifying structural abnormalities that are relevant to epilepsy, but MRI is more expensive and not immediately available in many regions. A first epileptic seizure is a common presentation, but little research has investigated the utility of MRI in this population, and in particular its additional value to that of CT, and its ability to predict seizure recurrence.
A prospective cohort study, based in a hospital-based First Seizure Clinic, was conducted. For 105 consecutive adults presenting with a first seizure, the frequency that MRI added value to CT was examined, and a logistic regression was performed to determine if clinical variables could predict when this would occur. The ability of neuroimaging to predict seizure recurrence was assessed using Kaplan-Meier curves.
MRI provided useful information additional to CT in 16.2% of patients, identifying six lesions missed by CT, and clarifying the nature of the CT lesion in 11, including 3 CT lesions which were shown to be artifactual. MRI provided additional information more often in patients who had a CT showing a lesion that required clarification (p=0.001), a partial-onset seizure (p=0.002), or a history of head-injury (p=0.009), and significantly less often in patients with a family history of epilepsy (p=0.046). If MRI was only offered to patients with a CT lesion that required clarification; or a partial-onset seizure; or a history of head-injury, then only 28.5% of patients would have received a MRI and 53% (16/30) of these MRIs would have added value to the CT. Of the patients not offered a MRI, only 1.3% (1/75) would have gained value from this test. Neuro-imaging did not predict seizure recurrence (p=0.78).
The results of this study identify a subgroup of patients in whom MRI has a high yield for adding value to CT, and indicates that MRI is not necessary in all patients presenting with a first seizure.
[Supported by: Department of Human Services, Victoria, Australia.]