Surgical treatment of drug-resistant focal epilepsy : selection, economic considerations and long-term outcomes

Epilepsy surgery can be an effective treatment for drug-resistant focal epilepsy, but requires careful selection of appropriate candidates to achieve optimal results and minimise the chance of adverse outcomes, including seizure recurrence. Long-term data on multimodal outcomes and a better appreciation of various factors influencing surgical suitability will help facilitate informed discussions between clinicians and prospective surgical candidates. This thesis includes a comprehensive analysis of a cohort of individuals who had epilepsy surgery at a tertiary neurosciences centre over the last 30 years, supplemented by data on individuals who completed presurgical evaluation at the same centre but did not proceed to surgical resection. An inability to localise the epileptogenic zone (EZ), multifocal epilepsy, or an individual choice not to have either intracranial electroencephalography or surgery were the most common reasons why people referred for presurgical evaluation did not proceed to a definitive resection. A predictive model based on demographic, imaging and electroclinical data was constructed to assist early discussions about surgical suitability. Those with normal MRI, extratemporal epilepsy and evidence of bilateral seizure onsets on video telemetry had an estimated 2.9% chance of proceeding to surgery, and people with a normal MRI brain invariably required intracranial EEG. Frontal lobe epilepsy surgery was safe and effective, with a long-term (median seven years) seizure freedom rate of 27%, and another 11% having auras only. Psychiatric comorbidity frequently improved postoperatively and paralleled seizure freedom. Focal MRI abnormality and fewer anti-seizure medications at the time of surgery could help predict a good outcome. In contrast, the site of resection and intracranial monitoring were not independently significant. Localisation of the EZ should rely on clinical features and multimodal investigation, as in our cohort frontal lobe semiology alone could correctly lateralise the EZ in only 77% and localise to a sublobar level in 52%. For those who completed presurgical evaluation but did not have surgery, under 5% had >12 months of seizure-freedom following the decision not to proceed, although a quarter had substantial improvement with further anti-seizure medication (ASM) or neurostimulation. Evaluation for epilepsy surgery was lengthy for individuals and costly for the public health system. Both duration and cost depended on what investigations were required and varied according to different routes through the presurgical pathway, especially the need for intracranial EEG. The median duration of evaluation was 29.7 months (IQR 18.6-44.1 months), with a median cost per person of £9,138 (IQR £6,984-£14,868). Approximately £123,500 was spent per additional person seizure-free.