(120 days)
The Ceribell Pocket EEG Device is intended to record and to present the EEG signals in visual and audible formats in real time. The visual and audible signals assist trained medical staff to make neurological diagnoses. The Pocket EEG Device is intended to be used in a professional healthcare facility environment.
Additionally, the EEG Recording Viewer Software component of the Pocket EEG Device incorporates a Seizure Detection component that is intended to mark previously acquired sections of EEG recordings in patients greater than or equal to 18 years of age that may correspond to electrographic seizures in order to assist qualified clinical practitioners in the assessment of EEG traces. The Seizure Detection component provides notifications to the user when detected seizure prevalence is "Frequent," "Abundant," or "Continuous," per the definitions of the American Clinical Neurophysiology Society Guideline 14. Notifications include an on-screen display on the Pocket EEG Device and the optional sending of an e-mail message to a clinician. Delays of up to several minutes can occur between the beginning of a seizure and when the Seizure Detection notifications will be shown to a user.
The Pocket EEG Device does not provide any diagnostic conclusion about the subject's condition and Seizure Detection notifications cannot be used as a substitute for real time monitoring of the underlying EEG by a trained expert.
The Ceribell Pocket EEG Device is a previously cleared EEG monitoring system which includes a portable 8-channel system that is being upgraded in this 510(k) notification K191301 to include a seizure detection module. The device connects to 10 patient electrodes (5 left, 5 right), which are used to form the 8 channels. The device may be used with any scalp EEG electrodes, and the system includes the following components:
- Pocket EEG Device: a portable, battery powered, 8-channel EEG monitoring system.
- Power adapter: 100-240 V AC power adapter used to charge the Pocket EEG Device.
- Micro-USB cable: cable used to connect Pocket EEG Device to power adapter for charging and to connect to a computer to transfer EEG recording files. When the Pocket EEG Device is connected to a power adapter of a computer, all EEG acquisition functions are automatically disabled.
- EEG Recording Viewer Software: EEG review software for viewing EEG recordings on a computer. The EEG Recording Viewer Software includes a Seizure Detection software module that assists qualified users in reviewing and annotating EEG by marking previously acquired sections of EEG that may correspond to electrographic seizures.
Here's a breakdown of the acceptance criteria and the study proving the device meets them, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
| Test Description | Acceptance Criteria | Reported Device Performance (PPA, 95% CI Lower Bound) | Conformance to PPA Acceptance Criteria | Reported Device Performance (FP/hr, 95% CI Upper Bound) | Conformance to FP/hr Acceptance Criteria |
|---|---|---|---|---|---|
| "Frequent" seizure activity notifications (≥10% seizure burden), PPA and FP/hr | Ceribell non-inferior to Persyst 13 with 10% non-inferiority margin | Not directly reported in table | N/A (non-inferiority not in table) | Not directly reported in table | N/A (non-inferiority not in table) |
| "Frequent" seizure activity notifications (≥10% seizure burden), PPA and FP/hr | PPA: 95% CI lower bound ≥ 70% | 90.00% | PASS | 0.351 | PASS |
| "Abundant" seizure activity notifications (≥50% seizure burden), PPA and FP/hr | FP/hr: 95% CI upper bound ≤ 0.446 | ||||
| "Abundant" seizure activity notifications (≥50% seizure burden), PPA and FP/hr | PPA: 95% CI lower bound ≥ 70% | 81.39% | PASS | 0.381 | PASS |
| "Continuous" seizure activity notifications (≥90% seizure burden), PPA and FP/hr | FP/hr: 95% CI upper bound ≤ 0.446 | ||||
| "Continuous" seizure activity notifications (≥90% seizure burden), PPA and FP/hr | PPA: 95% CI lower bound ≥ 70% | 78.26% | PASS | 0.272 | PASS |
| FP/hr: 95% CI upper bound ≤ 0.446 | |||||
| Seizure Burden output, L1-distance vs. reference standard | Ceribell non-inferior to Persyst 13 with 10% non-inferiority margin | Not directly reported in table | N/A | Not directly reported in table | N/A |
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: 60 subjects (30 from ICU dataset and 30 from EMU dataset).
- Data Provenance: Retrospective. Data was obtained from previous studies that retrospectively reviewed all adult EEGs from patients who underwent inpatient EEG monitoring over a period of time. Data from one hospital was used for ICU EEG data, and data from a second hospital was used for EMU EEG data. The study states the validation dataset was "from previous studies" and "retrospectively obtained from pre-existing EEG databases." The specific country of origin is not explicitly stated, but the submission is to the U.S. FDA, implying the data is likely from the U.S.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
- Number of Experts: An "expert panel" was used, with an agreement between "at least two neurologists required." The exact number of individual experts beyond "at least two" is not specified.
- Qualifications of Experts: EEG trained neurologists (physicians who have obtained fellowship training in epilepsy or neurophysiology). They were affiliated with "multiple different institutions."
4. Adjudication Method for the Test Set
- The ground truth was established by an "expert panel, with agreement between at least two neurologists required to determine the reference standard." This indicates an adjudication method where agreement between at least two experts was necessary. It suggests a form of consensus-based ground truth.
5. Was a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Done? If so, what was the effect size of how much human readers improve with AI vs without AI assistance?
- No, a multi-reader, multi-case (MRMC) comparative effectiveness study with human readers assisting the AI or vice-versa was not explicitly described in this document. The study compared the Ceribell device's performance to a predicate device (Persyst 13) and a consensus-based ground truth established by experts. The device's role is to "assist qualified clinical practitioners" but the study format does not evaluate the improvement of human readers with AI assistance.
6. If a Standalone (i.e. algorithm only without human-in-the loop performance) was done?
- Yes, a standalone performance evaluation of the Ceribell Seizure Detection module was done. The module was run on the validation dataset and its performance metrics (PPA and FP/hr) were calculated against the established reference standard. While the device's intended use is to "assist trained medical staff," the validation primarily assessed the algorithm's ability to detect seizures independently.
7. The Type of Ground Truth Used
- The ground truth used was expert consensus. It was established by "EEG trained neurologists who reviewed and annotated the EEG recordings for seizure episodes." Specifically, "Each seizure episode was manually annotated by an expert panel, with agreement between at least two neurologists required to determine the reference standard."
8. The Sample Size for the Training Set
- The sample size for the training set is not specified in the provided document. The document explicitly states that the validation dataset was "completely separate and independent from the data used to design and train the algorithm," and "No patients from the validation dataset were used for algorithm training or development," but it does not provide details on the training data itself.
9. How the Ground Truth for the Training Set Was Established
- The document does not provide details on how the ground truth for the training set was established. It only emphasizes that the validation dataset was separate from the training data.
§ 882.1400 Electroencephalograph.
(a)
Identification. An electroencephalograph is a device used to measure and record the electrical activity of the patient's brain obtained by placing two or more electrodes on the head.(b)
Classification. Class II (performance standards).