OBJECTIVES To evaluate the influence of clinical, investigative, and pathological factors on seizure remission after temporal lobectomy for medically intractable epilepsy associated with focal lesions other than hippocampal sclerosis. METHODS From a series of 234 consecutive "en bloc" temporal resections for medically intractable epilepsy performed between 1976 and 1995, neuropathological examination disclosed a focal lesion in 80. The preoperative clinical, neuropsychological, interictal EEG, and neuroimaging characteristics of these patients were assembled in a computerised database. The original neuropathological material was re-examined for lesion classification and completeness of removal. The presence of additional cortical dysplasia and mesial temporal sclerosis was also noted. Survival analysis was performed using Kaplan-Meier curves and actuarial statistics. Logistic regression analysis was used to establish the independent significance of the clinical variables. RESULTS The probability of achieving a 1 year seizure remission was 71% by 5 years of follow up. Factors predicting a poor outcome on multivariate analysis included the need for special schooling and a long duration of epilepsy. Generalised tonic-clonic seizures, interictal EEG discharges confined to the resected lobe, demonstration of the lesion preoperatively on CT, and complete histological resection of the lesion were not predictive of outcome. Neuropsychological tests correctly predicted outcome in left sided cases but apparently congruent findings in right sided resections were associated with a poor outcome. Pathological reclassification established the dysembryoplastic neuroepithelial tumour as the commonest neoplasm (87%) in this series, with a significantly better seizure outcome than for developmental lesions, such as focal cortical dysplasia. CONCLUSIONS The findings highlight the importance of dysembryoplastic neuroepithelial tumour in the pathogenesis of medically refractory lesional temporal lobe epilepsy and the prognostic significance of preoperative duration of epilepsy emphasises the need for early recognition and surgical treatment. Cognitive and behavioural dysfunction, however, is associated with a lower seizure remission rate, independent of duration of epilepsy.