Occipital epilepsy: spatial categorization and surgical management

Occipital resections for epilepsy are rare. Reasons for this are the relative infrequency of occipital epilepsy, difficulty in localizing epilepsy originating in the occipital lobe, imprecisely defined seizure outcome in patients treated with focal occipital resections in the MR imaging era, and concerns about producing visual deficits. The impact of lesion location on vision and seizure biology, the management decision-making process, and the outcomes following resection need elaboration.The authors studied 21 consecutive patients who underwent focal occipital resections for epilepsy at Cleveland Clinic Epilepsy Center over a 13-year period during which MR imaging was used. Demographics, imaging, and data relating to the epilepsy and its surgical management were collected. The collateral sulcus, the border between the medial surface and the lateral convexity, and the inferior temporal sulcus were used to subdivide the occipital lobe into medial, lateral, and basal zones. Lesions that did not involve most or all of the occipital lobe (sublobar) were spatially categorized into these zones. Visual function, semiology, and scalp electroencephalography were evaluated in relation to these spatial categories. Preresection and postresection visual function and seizure frequency were evaluated and compared. Lastly, an exhaustive review and discussion of the published literature on occipital resections for epilepsy was carried out.Five lesions were lobar and 16 were sublobar. Patients with medial or lobar lesions had a much greater likelihood of preoperative visual field defects. Those with basal or lateral lesions had a greater likelihood of having a visual aura preceding some or all of their seizures and a trend (not significant) toward having a concordant lateralized onset by scalp electroencephalography. Invasive recordings were used in 8 cases. All patients had lesions (malformations of cortical development, tumors, or gliosis) that were completely resected, as evaluated on postoperative MR imaging. At last follow-up, 17 patients (81%) were seizure free or had only occasional auras (Wieser Class 1 or 2). The remaining 4 patients (19%) had a worthwhile improvement in seizure control (Class 3 or 4). Of the patients for whom both pre- and postoperative visual testing data were available, 50% suffered no new visual deficits, and 17% each developed a new quadrantanopia or a hemianopia.Lesional occipital lobe epilepsy can be successfully managed with resection to obtain excellent seizure-free rates. Individually tailored resections (in lateral occipital lesions, for example) may help preserve intact vision in a subset of cases (38% in this series). Invasive recordings may further guide surgical decision-making as delineated by an algorithm generated by the authors. The authors' results suggest that the spatial location of the lesion correlates both with the semiology of the seizure and with the presence of visual deficit.