Earlier Identification of Candidates for Epilepsy Surgery with Machine Learning

Roughly one third of patients with epilepsy continue to have seizures despite treatment with antiseizure medications. Among these drug-resistant epilepsy (DRE) patients, those deemed good candidates through presurgical evaluation can have significant seizure improvement after surgical resection. Strategies to shorten the duration between DRE diagnosis and surgery are needed, as only 1% of DRE patients receive surgery within 2 years. In this multicenter, longitudinal cohort study, investigators sought to develop a clinically useful machine-learning algorithm that can accurately distinguish surgical from nonsurgical candidates before their scheduled neurology visit using information from prior neurology notes, electroencephalogram and MRI rep…