Monitoring the Burden of Seizures and Highly Epileptiform Patterns in Critical Care with a Novel Machine Learning Method

Abstract

Introduction Current electroencephalography (EEG) practice relies on interpretation by expert neurologists, which introduces diagnostic and therapeutic delays that can impact patients’ clinical outcomes. As EEG practice expands, these experts are becoming increasingly limited resources. A highly sensitive and specific automated seizure detection system would streamline practice and expedite appropriate management for patients with possible nonconvulsive seizures. We aimed to test the performance of a recently FDA-cleared machine learning method (Claritγ, Ceribell Inc.) that measures the burden of seizure activity in real time and generates bedside alerts for possible status epilepticus (SE). Methods We retrospectively identified adult patients (n = 353) who underwent evaluation of possible seizures with Rapid Response EEG system (Rapid-EEG, Ceribell Inc.). Automated detection of seizure activity and seizure burden throughout a recording (calculated as the percentage of ten-second epochs with seizure activity in any 5-min EEG segment) was performed with Claritγ, and various thresholds of seizure burden were tested (≥ 10% indicating ≥ 30 s of seizure activity in the last 5 min, ≥ 50% indicating ≥ 2.5 min of seizure activity, and ≥ 90% indicating ≥ 4.5 min of seizure activity and triggering a SE alert). The sensitivity and specificity of Claritγ’s real-time seizure burden measurements and SE alerts were compared to the majority consensus of at least two expert neurologists. Results Majority consensus of neurologists labeled the 353 EEGs as normal or slow activity (n = 249), highly epileptiform patterns (HEP, n = 87), or seizures [n = 17, nine longer than 5 min (e.g., SE), and eight shorter than 5 min]. The algorithm generated a SE alert (≥ 90% seizure burden) with 100% sensitivity and 93% specificity. The sensitivity and specificity of various thresholds for seizure burden during EEG recordings for detecting patients with seizures were 100% and 82% for ≥ 50% seizure burden and 88% and 60% for ≥ 10% seizure burden. Of the 179 EEG recordings in which the algorithm detected no seizures, seizures were identified by the expert reviewers in only two cases, indicating a negative predictive value of 99%. Discussion Claritγ detected SE events with high sensitivity and specificity, and it demonstrated a high negative predictive value for distinguishing nonepileptiform activity from seizure and highly epileptiform activity. Conclusions Ruling out seizures accurately in a large proportion of cases can help prevent unnecessary or aggressive over-treatment in critical care settings, where empiric treatment with antiseizure medications is currently prevalent. Claritγ’s high sensitivity for SE and high negative predictive value for cases without epileptiform activity make it a useful tool for triaging treatment and the need for urgent neurological consultation.

Noninvasive mobile EEG as a tool for seizure monitoring and management: A systematic review

“…They found that epileptologists, without automated data annotation, had a lower sensitivity (61%) but better false‐alarm rate (0.002/h) compared to the automated seizure‐detection algorithm (with no epileptologist involvement) that achieved a sensitivity of 90% and a false‐alarm rate of 0.087/h. Finally, Karmousi et al 42 evaluated a machine learning method to automatically estimate “seizure burden,” defined as the number of 10 s epochs with seizure activity in any 5 min period, with thresholds for low, medium, and high seizure burden (seizure activity in 10%, 50%, and 90% of epochs); detection of high seizure burden was used to generate a “status epilepticus” alert. EEG data were collected using the Ceribell in patients in the ICU.…”

“…Two of the studies presented some quality concerns. Frankel and colleagues 27 used a nonbalanced number of events for the evaluation of the diagnostic accuracy of manual seizure detection (31 epochs with ictal events and 83 nonictal), whereas Kamousi et al 42 pointed out that their cohort contained a relatively low number of patients with high seizure burden (9 of 353 EEG studies).…”

The Past, Present, and Future of Tele-EEG

“…Additionally, emerging hardware and software solutions that allow for rapid-response EEG can reduce time to diagnosis and definitive therapy as well as length of hospital stay for inpatients with suspected seizures. 82 Technological advances can also support more efficient deployment of aEEG in the nonacute setting.…”

“…The future of tele-EEG is exciting in that spike and seizure detection algorithms can directly scale these diagnostic services 82 by reducing constraints imposed by personnel shortages and can also help us study the clinical questions that may validate and grow the current models of utilization. 47 While subtle seizures can go undetected by these software, the algorithms have become astonishingly accurate relative to board-certified clinical neurophysiologists.…”

Rapid Response EEG: Current State and Future Directions

“…A retrospective cohort study reported high performance of clarity in 353 adult patients in the emergency departments and critical care units across six academic and community hospitals who underwent evaluation for altered mental status and NCSE using the RRLM-EEG device [ 30 ]. When compared with neurologists, authors reported high sensitivity (100%) and specificity (93%) when evaluating EEG with > 90% seizure burden in status epilepticus.…”