A Retrospective Study of Artificial Intelligence Derived Seizure Burden Scores on Outcomes in Patients Undergoing Rapid Electroencephalography
Abstract number :
2.137
Submission category :
3. Neurophysiology / 3C. Other Clinical EEG
Year :
2024
Submission ID :
382
Source :
www.aesnet.org
Presentation date :
12/8/2024 12:00:00 AM
Published date :
Authors :
Presenting Author: Greta Brown, BS – College of Osteopathic Medicine of the Pacific-Northwest, Western University of Health Sciences
Alyx Lesko, BS – Providence Brain and Spine Institute
Horia Marginean, MD, MS – Providence Brain and Spine Institute
Weston Anderson, BA – Providence Brain and Spine Institute
Evan Fertig, MD – Providence Brain and Spine Institute
Rationale: Rapid limited montage EEG (rEEG) with artificial intelligence (AI) monitoring has emerged as an alternative to formal video-electroencephalography (VEEG). rEEG-AI provides continuous seizure burden scores to guide management1. Studies have shown seizure burden in status epilepticus (SE) influences outcomes2. We evaluated outcomes among patients monitored with Ceribell rEEG-Clarity AI. We hypothesized patients with high seizure burden scores would fare worse.
Methods: Retrospective data from electronic medical records for patients >18 years old with rEEG-AI for any indication between Nov 1, 2021-Apr 30, 2023 admitted to one of four Portland Oregon medical facilities, were included. Patients with anoxic ischemic injury due to cardiac arrest were excluded. Outcome measures:
· Discharge disposition.
· Modified Rankin scale (mRS) at discharge categorized as “good” (0-2 or return to pre-admission mRS) and “bad”.
· Glasgow coma scale.
· Hospital length of stay (LOS).
·Intensive care unit LOS.
· Change in antiseizure medications prescribed at discharge (ASMd).
mRS was inferred from the chart based on published methodology3. We used multivariate regression models to compare maximum seizure burden score (AI-MSB), continuous and categorized, to outcomes. Models were adjusted for etiology of presentation and history of epilepsy.
Results: One-hundred and sixty-six patients were included (Table 1). MSB (median 3 (0-27) was categorized into three groups: 0-10 (59.0%), 10-90 (33.2%), and 90-100 (7.8%). In univariate and multivariate analysis, MSB was not significantly correlated to outcomes in our cohort (Table 2).
Conclusions: We did not find a relationship between AI-MSB and clinical outcomes as demonstrated by previous studies of status epilepticus in children. There are multiple possibilities why this relationship was not found. Our study sample was small, adult only and included AI-MSB with and without seizures and SE. In addition, Clarity seizure burden scores are complex including multiple variables not limited to epileptiform activity.
References:
< !1. Kamousi, B., Karunakaran, S., Gururangan, K., Markert, M., Decker, B., Khankhanian, P., Mainardi, L., Quinn, J., Woo, R., & Parvizi, J. (2021). Monitoring the Burden of Seizures and Highly Epileptiform Patterns in Critical Care with a Novel Machine Learning Method. Neurocritical Care, 34(3), 908–917.
< !2. Payne ET, Zhao XY, Frndova H, McBain K, Sharma R, Hutchison JS, Hahn CD. Seizure burden is independently associated with short term outcome in critically ill children. Brain. 2014 May;137(Pt 5):1429-38. doi: 10.1093/brain/awu042. Epub 2014 Mar 4. PMID: 24595203; PMCID: PMC3999716.
Neurophysiology