Predicting Post-Traumatic Epilepsy Using Admission Electroencephalography after Severe Traumatic Brain Injury
Abstract number :
1.462
Submission category :
2. Translational Research / 2C. Biomarkers
Year :
2022
Submission ID :
2232899
Source :
www.aesnet.org
Presentation date :
12/3/2022 12:00:00 PM
Published date :
Nov 22, 2022, 05:28 AM
Authors :
Matthew Pease, MD – Memorial Sloan Kettering; Jonathan Elmer, MD – University of Pittsburgh; Arka Mallela, MD – University of Pittsburgh; Juan Ruiz-Rodriguez, MD – University of Washington; Daniel Sexton, MD – Duke University; Nirav Barot, MD, MS – University of Pittsburgh; Jorge Gonzalez-Martinez, MD, PhD – University of Pittsburgh; Lori Shutter, MD – University of Pittsburgh; David Okonkwo, MD, PhD – University of Pittsburgh; James Castellano, MD, PhD – University of Pittsburgh
This is a Late Breaking abstract
Rationale: Post-traumatic epilepsy (PTE) develops in as many as one-third of severe traumatic brain injury (TBI) patients, often years after injury. Analysis of early electroencephalography (EEG) features, both standardized visual (viEEG) interpretation and quantitative EEG (qEEG) analysis, may aid early identification of patients at high risk for PTE.
Methods: We performed a case-control study using a prospective database of severe TBI patients treated at a single center from 2011 through 2018. We identified all patients who survived two years post-injury and matched patients with PTE to those without using age and Glasgow Coma Scale score at admission. A trained neuropsychologist recorded outcomes at one-year using the Expanded Glasgow Outcomes Scale (GOSE). All patients underwent continuous EEG for 3 to 5 days. A board-certified epileptologist, blinded to the outcome, described viEEG features using standardized descriptions. We extracted fourteen qEEG features from an early 5-minute epoch, described them with qualitative statistics, and developed a random forest model to predict long-term risk of PTE.
Results: We identified 27 patients with and 23 without PTE. GOSE were similar at one-year (p=0.92). The median time to onset of PTE was 7.2 months post-trauma (interquartile range: 2.2-22.2 months). viEEG findings of background alpha activity was protective against PTE (p=0.04). On qEEG analysis, the PTE cohort had higher spectral power in the delta frequencies, more power variance in the delta and theta frequencies, higher mean amplitude, and higher peak envelope (all p< 0.05). Our model, combining qEEG, viEEG, and clinical features, had an area under the curve (AUC) of 0.81 (0.69-0.93). We implemented Shapley Additive Explanations, a common post-hoc analysis approach used to explain machine learning models. Delta variance, background alpha activity, and delta mean power were the three most important features for model performance.
Translational Research