For cavernous haemangiomas, it is the aim of surgical treatment to control epilepsy and eliminate potential sources of intracerebral haematomas. In the following investigation, it was attempted to find indicators for seizure freedom after surgery. Success of therapy was assessed according to three patterns of classification. Thirty patients underwent tailored resection based on findings from preoperative investigations and intraoperative electrocorticography. Follow-up averaged 4 years. Lesionectomy, extended lesionectomy, and modified lobe resection were carried out in 13, 11, and six patients, respectively. For all procedures, including microsurgical lesionectomy, the firm gliotic layer unequivocally differed in colour and consistency from normal brain and was removed. Further tissue resection was carried out only if the electrocortical course suggested persistent spike activity around the resection cavity or if presurgical MRI evaluation (e.g. hippocampal atrophy) or electrophysiology also pointed to pathology distant from the lesion. Of the patients, 53.3% became completely seizure-free (Engel I), and one additional patient had only occasional isolated auras. Dramatic reductions in seizure frequency and severity were exhibited by 26.7%. Outcome in respect to seizure control was not associated with resection procedure, comparing pure lesionectomy with lesionectomy plus cortectomy. In the group of patients with epilepsy surgery, those with hippocampectomy had significantly better outcome than those without. Important prognostic factors were early operation after seizure manifestation (91.7% operated upon within 2 years of seizure onset became seizure-free). Another prognostic factor was unifocal seizure onset (bilateral or multifocal seizure onset was found in care of the ten patients with unfavourable outcome). None of the four patients harbouring multiple cavernomas became seizure-free after resection of one lesion, which was believed to be mostly attributable to the epileptic focus that was removed.