encevis (2.1)

encevis (2.1) is an advanced EEG analysis software designed to aid clinicians by providing comprehensive review tools and automated detection of critical neurological events such as seizures, spikes, and electrographic status epilepticus in adult patients. The system employs various algorithmic approaches, including AI-enhanced seizure detection and quantitative EEG analysis, to process scalp EEG data for improved monitoring and assessment, particularly in acute care settings.

encevis (2.1) is a software package designed for qualified medical practitioners to review, monitor, and analyze electroencephalogram (EEG) recordings obtained via scalp electrodes using the standard 10/20 system. It offers a suite of integrated modalities including an EEG-viewer for data visualization and basic analysis (filtering, scaling, montage adjustments), artifact reduction (PureEEG) for EMG and electrode artifacts, and automated detection components. These include seizure detection (NeuroTrend) for adult electrographic seizures, spike detection (EpiSpike) for adult interictal events, and a specialized component (encevis acute care) for detecting seizures and electrographic status epilepticus (ESE) in adult acute care patients, which also calculates seizure burden. The system further provides pattern detection (rhythmic/periodic patterns, burst suppression based on ACNS terminology), amplitude-integrated EEG (aEEG) for brain state monitoring, quantitative EEG measures like frequency bands, and spectrogram visualization using wavelet analysis. encevis can operate in both offline mode with stored EEG files and online mode processing real-time streaming data, providing on-screen notifications or alerts via external interfaces, though it emphasizes that its outputs are aids and not standalone diagnostic conclusions, except for the ESE component's diagnostic output in acute care. Performance testing demonstrated varying results across its modules when compared to other devices or expert consensus. The general seizure detection (NeuroTrend) showed a higher positive percentage agreement (PPA) of 93.8% compared to Persyst 12's 77.3%, but also a higher negative disagreement rate (NDR) of 33.7 false detections per 24 hours versus 10.5. Similarly, the acute care seizure detection component had a PPA of 71.6% (vs. Persyst's 41.5%) with an NDR of 2.0/hour (vs. 0.26/hour), indicating a more sensitive but less specific operating point, especially effective for longer seizures. For ESE detection, encevis achieved a PPA of 82.6% and NPA of 91.4%. Seizure burden calculations showed good agreement for various thresholds (e.g., >10% hourly burden PPA 86.8%, >50% short-term burden PPA 88.6%). Spike detection PPA was significantly higher at 84.81% (vs. Persyst's 8.7%). Artifact reduction was found to be non-inferior to Persyst. Rhythmic/periodic pattern detection, aEEG, frequency band analysis, burst suppression detection, and spectrogram functionalities also demonstrated acceptable performance against reference standards or comparable devices.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. Given the trade-off between higher sensitivity and higher false positive rates in seizure detection for both NeuroTrend and acute care components, what specific clinical scenarios or patient populations would benefit most from this operating point, and where might it be less optimal?; How does the performance of the encevis PureEEG artifact reduction specifically impact the accuracy of downstream analyses like seizure, spike, and electrographic status epilepticus detection, particularly in noisy ICU environments?; What is the typical notification delay for critical events like electrographic status epilepticus detection in online mode when integrated with various third-party EEG systems, and how does this compare to the potential "several minutes" delay mentioned?; Could the AI model integrated into the encevis 2.1 seizure detection algorithm be further trained or adapted for specific seizure semiologies or pediatric populations, given its current adult-only indication for seizure and spike detection?; For the encevis acute care component, how does the diagnostic output for electrographic status epilepticus and seizure burden directly influence treatment decisions in acute-care environments, and are there studies validating changes in patient outcomes based on its use?