Aims To ascertain the relevance of structural abnormalities in epilepsy surgery.

Methods We reviewed the structural findings and outcomes of patients who had been discussed at the multidisciplinary team meeting and were subsequently operated or rejected. Outcome was determined 2 years after surgery using Engel's classification.

Results 443 patients were included, 307 (69%) had a structural abnormality on MRI. The most common abnormality was hippocampal sclerosis. Other abnormalities included cavernomas, dysplasia, tumours and infarcts. 203 patients were operated, 240 were not operated. The presence of a structural abnormality was clearly associated with the selection for surgical treatment. 182 of 203 operated patients had a structural abnormality. 125 of 240 nonoperated patients had a structural abnormality (p=4.37×10⁻¹¹). However, the presence of a structural abnormality was not associated with outcome of surgery. In the surgery group 154 had temporal lobectomies or selective amygdalohippocampectomies, 36 had lesionectomies, 13 had other operations. 101 of 182 operated patients with a structural abnormality became seizure free. Eight of 21 operated patients without a structural abnormality became seizure free (p=0.13).

Conclusion Selection of patients for surgery but not outcome of surgery was associated with structural abnormalities.