Abstracts

RATIONALE: To compare language lateralization using BOLD functional magnetic resonance imaging (fMRI) with results of the intracarotid amobarbital procedure (Wada test) in a large number of patients with epilepsy, including a substantial number of patients with atypical language lateralization in the Wada test.
METHODS: We investigated 100 patients with localization-related epilepsy differing widely in age and IQ (52 f; age median 31, range 12-60 years; IQ median 89, range 46-130). During presurgical assessment for epilepsy surgery all patients had a Wada test (cerebral angiography + hemispheric anesthetization + language testing of the awake hemisphere). Using a standard 1.5 T Siemens Symphony scanner, a standard EPI sequence (16 axial 4mm thick slices, FOV 192 mm, 64x64 matrix, TR 1600 msec, TE 50 msec), and the standard on-line statistical postprocessing software of the scanner, we contrasted images sampled during 10 episodes of word generation with images sampled during 10 episodes of a low level rest condition in all patients individually. The resulting statistical maps were judged visually as [ssquote]typical[ssquote], [ssquote]atypical[ssquote], or [ssquote]artefactual[ssquote].
RESULTS: By Wada test standards 70 patients showed a typical, 30 an atypical language lateralization. The acquisition of fMRI data was easy and quick (15 min/patient). The assessment of fMRI-results was reproducible (interrater kappa = 0.86). FMRI results were judged artefactual due to movement in 6 patients. Bilateral fMRI activation was the most frequent finding (86%). Asymmetric activations, however, were the basis of the classification (92%). FMRI provided typical and atypical results in 85 of the remaining 94 individual cases in concordance with the Wada test. There were discordant results between fMRI and Wada test in 9 patients. Atypical language lateralization was associated with leftsided extratemporal lobe epilepsy and with a younger age at an early precipitating injury.
CONCLUSIONS: There was a high concordance lateralizing language between both tests. 10% discordant findings resulted from bilateral fMRI activations, most likely representing the non-dichotomous nature of a brain network for language. In a sequence of diagnostic tests, fMRI of language lateralization can only serve as a filter for the Wada test and thus potentially reduce the number of patients undergoing invasive investigations prior to epilepsy surgery, when combined with clinical information.
[Supported by: Society for Epilepsy Research]