[Multimodal imaging in epilepsy surgery. Morphological patterns of the cerebral cortex; modeling the limits of normality].

Precise localization of the epileptogenic zone, which is mandatory for effective surgery, relies on integration of multimodal information (clinical, electrical, magnetic, functional, etc.) within a common framework--a 3D morphological description of the cerebral cortex. Generation of this anatomical framework in individual patients is hindered by the extreme morphological variability of the human cerebral cortex. In order to find a technical solution to this clinical problem, we have developed a theoretical model of cerebral cortex sulcation, and mathematical tools allowing us to automatically identify the sulcal roots. A set of volunteers served to train a neuronal network ("Anatomist") that is now able to evaluate the morphological parameters of each silcus and to compare them with values obtained in normal subjects. This approach has been used to study the influence of handedness, gender, etc., in a population of normal subjects. It has also been used to detect specific abnormal patterns in genetic disorders like Turner's syndrome. Between September 1999 and July 2005, in the stereotactic and functional neurosurgery department of Timone University Hospital in Marseilles, we studied 103 epileptic patients with implanted deep electrodes (SEEG). Thirty-six patients had frontal lobe epilepsy, of whom 26 were considered MRI-negative by experienced neuroradiologists. In 12 patients, SEEG allowed us to precisely define the epileptogenic zone and to perform successful cortectomy. Follow-up is now longer than one year. These patients' cerebral cortex gyration patterns were then studied with the Anatomist program. The patients' mean age was 25 years, and mean age at epilepsy onset was 5 years. Anatomist was used to determine which sulci were morphologically abnormal. The topography of the epileptogenic zone, as determined by SEEG investigation and confirmed by successful cortectomy, was compared with the topology of the abnormal gyration. At last follow-up, 65% of the patients were in Engel class I, i.e. seizure-free. Abnormal sulcations were found in the epileptogenic zone in 9 (75%) of the 12 patients. Systematic analysis of normal gyration patterns in MRI-negative epileptic patients allowed us to detect subtle abnormalities that turned out to be reliable markers of the epileptogenic zone. Such markers may be of major value for defining the topology of the epileptogenic zone, in good agreement with information obtained with different modalities, and may significantly increase the chances of successful epilepsy surgery.