Curved reconstructions versus three-dimensional surface rendering in the demonstration of cortical lesions in patients with extratemporal epilepsy

RATIONALE AND OBJECTIVES. TO compare the visibility and localization of extratemporal cortical lesions in extratemporal epilepsy by using curved reconstruction (CR) and three-dimensional surface rendering (3D SR) of 3D-acquired MR images and to study the degree of confidence with which localizations are made, particularly at the gyral level.

METHODS. Twenty patients with extratemporal epilepsy, based on seizure symptomatology and/or scalp electroencephalographic registrations, with an extratemporal structural lesion on conventional MR imaging, were selected for this study by a neuroradiologist with extensive experience in the assessment of epilepsy patients, Transverse T2 spin-echo, coronal fluid-attenuated inversion recovery, and transverse 3D-acquired/two-dimensionally reconstructed T1 MR images were used for the selection. A second neuroradiologist (observer 1) and a radiology resident (observer 2) assessed CR and 3D SR in random order. Both observers were masked to all patient data. The subjective visibility of lesions and gyral location were scored. The interobserver agreements for lesion visibility and localization and for degree of confidence were compared for CR and 3D SR.

RESULTS. For both observers, the lesion was visible in 55% of 3D SRs and 95% of CRs. The proportion with "very clearly visible" lesions on 3D SR was 19% (4/20) according to observer 1 and 30% (6/20) according to observer 2. For CR, this proportion was substantially higher: 55% for both observers. This difference was significant for observer 1 but not for observer 2, The interobserver agreement was high for both methods. Agreement on gyral localization was 28% for CR and 40% for 3D SR. The percentage of similar confidence scores for the same gyral localization and for gyral localization with a maximum difference of one gyrus between the observers did not differ significantly for CR or 3D SR. The observers were more often confident in agreed cases in CR and moderately confident in 3D SR,

CONCLUSIONS. These results suggest that CRs of the brain surface are superior to 3D SR for the visualization of extratemporal cortical lesions in patients with drug-resistant extratemporal epilepsy, If lesions are seen, no significant difference was found between the two techniques for localization; however, the degree of confidence appears higher for CR at the gyral level.