MRI, EEG outperform CT for detecting refractory epilepsy

Combining neuroimaging and neurophysiological methods can help improve the early diagnosis of refractory epilepsy, a recent study has found.

Researchers retrospectively assessed 39 patients with refractory epilepsy who had undergone noninvasive examinations, such as computed tomography (CT), magnetic resonance imaging (MR), and electroencephalogram (EEG). The outcome of interest was the correlation among these tests.

In terms of aetiology, the most common manifestations of refractory epilepsy were hippocampal sclerosis (n=12), brain malformations (n=6), and cerebrovascular accidents (n=3).

Of the participants, 74.36 percent (n=29) showed positive MRI findings. MRI shared a strong correlation with aetiology (R, 1) and was able to detect 100 percent of all positive cases aetiologically as well as correctly identify all manifestations. EEG likewise had a high rate of agreement with the aetiology-based diagnosis (93.10 percent), while CT did not (41.38 percent).

Compared with CT, MRI showed significantly better performance in the detection of epileptogenic foci (p<0.01). However, no such difference was reported when MRI was compared with EEG (p>0.05). Moreover, the ratios of MRI/CT and CT/EEG were significantly different to the aetiology-based diagnoses (p<0.05), while the EEG/MRI ratio was not (p>0.05).

The findings confirmed that MRI remains one of the most sensitive methods by which refractory epilepsy can be diagnosed, matching its aetiological diagnosis nearly perfectly. The same could be said for EEG, which did not differ from MRI statistically, though some discrepancies vs aetiology were detected. In contrast, CT seems to be underpowered.

“MRI is highly sensitive to the aetiological diagnosis of refractory epilepsy and EEG is sensitive to localization of epileptogenic foci; the two are highly correlated. In clinical practice, early diagnosis of refractory epilepsy combined with neuroimaging and neurophysiological methods can improve the accuracy of diagnosis,” the researchers said.