High-resolution imaging techniques can demonstrate anatomic alterations in most patients identified as candidates for surgical treatment of their partial epilepsy. The demonstration of an anatomic lesion is only one step in the presurgical diagnostic procedure, which includes video-EEG and, when necessary, video-stereo-EEG recordings of seizures. A review of the literature shows that the simple removal of the magnetic resonance imaging (MRI)-evident lesion ("lesion-ectomy") reduces but does not completely suppress seizures in a large percentage of patients, especially those with neuronal migration disorders. This phenomenon could, at least in part, be explained by preliminary data (in 33 patients) showing that less than 20% of seizures correspond to a well-localized, intralesional discharge in about 40% of stereo-EEG-investigated patients with at least one intralesional electrode. The authors illustrate some anatomo-electroclinical examples of intraindividual variability of the ictal symptomatology, raising the problem of the decision about the extent of the surgical removal. Recent histologic and immunohistochemical studies have demonstrated several kinds of structural alterations in the stereo-EEG-defined epileptogenic zone, not always overlapping with the MRI-visible lesion. This aspect can further explain some failures of MRI-guided lesionectomies. That relationships between "lesions" and epileptogenic zones may be variable is also suggested by reports of patients who present with multiple lesions (i.e., cavernous angiomas, Bourneville syndrome) and are cured by removal of only one of them.