Tailored multilobar disconnective epilepsy surgery in the posterior quadrant.

Objective: Surgical treatment of drug-resistant epilepsy originating from the posterior quadrant (PQ) of the brain often requires large multilobar resections, and disconnective techniques have been advocated to limit the risks associated with extensive tissue removal. Few previous studies have described a tailored temporoparietooccipital (TPO) disconnective approach; only small series with short postoperative follow-ups have been reported. The aim of the present study was to present a tailored approach to multilobar PQ disconnections (MPQDs) for epilepsy and to provide details about selection of patients, presurgical investigations, surgical technique, treatment safety profile, and seizure and cognitive outcome in a large, single-center series of patients with a long-term follow-up.

Methods: In this retrospective longitudinal study, the authors searched their prospectively collected database for patients who underwent MPQD for drug-resistant epilepsy in the period of 2005-2017. Tailored MPQDs were a posteriori grouped as follows: type I (classic full TPO disconnection), type II (partial TPO disconnection), type III (full temporooccipital [TO] disconnection), and type IV (partial TO disconnection), according to the disconnection plane in the occipitoparietal area. A bivariate statistical analysis was carried out to identify possible predictors of seizure outcome (Engel class I vs classes II-IV) among several presurgical, surgical, and postsurgical variables. Preoperative and postoperative cognitive profiles were also collected and evaluated.

Results: Forty-two consecutive patients (29 males, 24 children) met the inclusion criteria. According to the presurgical evaluation (including stereo-electroencephalography in 13 cases), 12 (28.6%), 24 (57.1%), 2 (4.8%), and 4 (9.5%) patients received a type I, II, III, or IV MPQD, respectively. After a mean follow-up of 80.6 months, 76.2% patients were in Engel class I at last contact; at 6 months and 2 and 5 years postoperatively, Engel class I was recorded in 80.9%, 74.5%, and 73.5% of cases, respectively. Factors significantly associated with seizure freedom were the occipital pattern of seizure semiology and the absence of bilateral interictal epileptiform abnormalities at the EEG (p = 0.02). Severe complications occurred in 4.8% of the patients. The available neuropsychological data revealed postsurgical improvement in verbal domains, whereas nonunivocal outcomes were recorded in the other functions.

Conclusions: The presented data indicate that the use of careful anatomo-electro-clinical criteria in the presurgical evaluation allows for customizing the extent of surgical disconnections in PQ epilepsies, with excellent results on seizures and an acceptable safety profile.