The NightWatch+ US is a prescription only device that is indicated for use as an adjunct to seizure monitoring of children age 4 till 16 diagnosed with epilepsy having Nocturnal Epileptic Major Motor Seizures which includes tonic-clonic (TC), tonic (if clustered or prolonged >30 seconds), hyperkinetic and TC-like seizures, in home or residential facilities during periods of rest. The Sensor of the device is worn on the upper arm and measures heart rate and motion data to detect patterns that may be associated with nocturnal epileptic motor seizures in patients with epilepsy. When a seizure event is detected by the Sensor of the NightWatch+ US, it sends a command to the paired wireless alarm station of the NightWatch+ US that is programmed to initiate an alarm to a designated caregiver. The system records and stores data from seizure events. The data can be viewed by the user in a cloud based data portal. The NightWatch+ US is not intended to diagnose specific seizure types.

NightWatch+ US consists of a sensor worn during sleep on the biceps of the upper arm and an alarm station. The sensor consists of a heart rate sensor using PPG (photoplethysmography), a ACC (Accelerometry) movement sensor, and a microprocessor. The microprocessor processes the data from the sensors using a detection algorithm which detects if the sensor readings match pre-programmed parameters that are associated with nocturnal epileptic major motor seizures. When a nocturnal epileptic major motor seizures is detected, the seizure alarm is triggered and transferred from the sensor to the accompanying alarm station which alarms the caregiver by a sound and a blinking LED light. The seizure alarms data can be transferred from the alarm station, using an ethernet connection, to a database in the cloud to be viewed a web-based interface called NightWatch Portal to be able to monitor seizure frequency overtime.

The provided text describes the acceptance criteria and a study proving the device meets these criteria. However, it explicitly states that "performance goals or acceptance criteria for other endpoints were not defined" beyond an 80% sensitivity estimate for the study itself. This means that while the study evaluated performance metrics like sensitivity and false alarm rates, these were reported as results of the study, not as pre-defined acceptance criteria the device needed to meet for regulatory clearance. The document focuses on demonstrating substantial equivalence to a predicate device, arguing that the reported performance is comparable.

Therefore, the table of "Acceptance Criteria" will reflect the reported performance values from the study, as these are the figures the FDA appears to have accepted for clearance based on the substantial equivalence argument, rather than explicitly stated pre-market acceptance thresholds.

Acceptance Criteria and Device Performance

Study Details

1. Sample Size and Data Provenance:

   • Test Set Sample Size: 53 children aged 4-16 years.
   • Data Provenance: The study by van Westrhenen et al. 2023 was a "phase 4, multicenter, prospective, video-controlled, in-home study." The specific country of origin is not explicitly stated, but the company (LivAssured BV) is based in the Netherlands, and the lead author's affiliation (van Westrhenen) is often associated with Dutch institutions, suggesting data primarily from Europe.

2. Number of Experts and Qualifications for Ground Truth:

   • Number of Experts: Two principal investigators (R.D.T. and R.H.C.L.) consulted for final decisions in case of discrepancies or doubt in video annotations. Their specific qualifications (e.g., years of experience, direct specialty) are not detailed beyond being "principal investigators" in a study focused on epilepsy, implying clinical expertise in neurology/epileptology.
   • Other Reviewers: "Trained trial nurses" annotated events and retrospectively analyzed video tracings.

3. Adjudication Method for the Test Set:

   • The primary method involved "Trained trial nurses" annotating all events (NightWatch alarms, video alarms, and caregivers' seizure diary) while blinded to alarm type and sensor data.
   • Adjudication Process: In cases of discrepancies between nurses' annotations or doubt, the trial nurses consulted one of the two principal investigators for a final decision.
   • Quality Control: The principal investigators double-checked a random sample of 5% of the annotations. Additionally, trained trial nurses fully screened the video of 5% of all nights for missed seizures.

4. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

   • No, a formal MRMC comparative effectiveness study was explicitly mentioned or described wherein human readers' performance with AI assistance was compared to their performance without AI assistance. The study focuses on evaluating the device's standalone performance and demonstrating its functionality as an adjunct to seizure monitoring.

5. Standalone Performance:

   • Yes, the study primarily assessed the standalone performance of the NightWatch+ US algorithm in detecting nocturnal epileptic major motor seizures. The reported sensitivity and false alarm rates reflect the algorithm's performance in detecting events based on physiological signals (heart rate and motion data) and alerting caregivers.

6. Type of Ground Truth Used:

   • Video Monitoring / Expert Consensus: The ground truth was established through extensive video monitoring of patients. Trained trial nurses annotated events from these video recordings. In cases of disagreement or doubt, principal investigators provided a final decision, effectively establishing an expert consensus based on video evidence. The document states, "Video is an equally robust reference standard to Video-EEG provided the whole dataset is reviewed for inferring the reference standard as recognized by the International League Against Epilepsy (ILAE)."
   • Caregiver Diaries: Caregivers' seizure diaries were also part of the input scrutinized alongside video, but video was the definitive reference.

7. Sample Size for Training Set:

   • The provided 510(k) summary does not specify the sample size or characteristics of the training set used for the NightWatch+ US algorithm. It states that the "seizure detection method, algorithms and components used for seizure detection are identical" to the predecessor NightWatch and NightWatch+, implying development and training occurred prior to this specific clearance study, but details about that initial training data are not found in this document.

8. How Ground Truth for Training Set Was Established:

   • As the training set details are not provided in this regulatory document, the method for establishing ground truth for any potential training data is also not described. It's implied that similar clinical data (possibly from the predecessor device's accumulated experience and studies) would have been used for algorithm development and refinement, which typically involves expert review of clinical events.

Ask a specific question about this device

EpiMonitor is a prescription only medical device system composed of a wearable device "EmbracePlus" and paired mobile software application "EpiMonitor" intended as an adjunct to seizure monitoring in adults and children aged 6 and up in a home environment or healthcare facilities. The device is worn on the wrist and senses Electrodermal Activity (EDA) and motion data to detect patterns that may be associated with either primary or secondary generalized tonic clonic seizures in patients with epilepsy or at risk of having epilepsy. When a seizure event is detected, the wearable device component of EpiMonitor sends a command to a paired mobile device where the EpiMonitor App is programmed to initiate an alert to a designated caregiver. The EpiMonitor app incorporates additional detection sensitivity modes, "high" for use during periods of rest or sleeping or "low" for use during periods of low-intensity activity, in order to reduce false alarm incidents.

EpiMonitor records, stores and transmits accelerometer, EDA, peripheral skin temperature and activity data for subsequent retrospective review by a trained healthcare professional via a cloud-based software.

The EpiMonitor system consists of a wearable device and mobile application:

• A wearable medical device called EmbracePlus,
• A mobile application running on smartphones called "EpiMonitor"

The EmbracePlus is worn on the user's wrist and continuously collects raw data via specific sensors, these data are continuously analyzed by an on-board algorithm (EpiAlgo 2.1), which assesses the physiological data and determines if the user may be undergoing a generalized tonic-clonic seizure (GTCS). The EpiAlgo has been validated through testing, using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units.

When a likely GTCS is detected, EmbracePlus sends, via Bluetooth Low Energy, a message to the EpiMonitor app. The EpiMonitor app communicates to the Empatica Cloud which initiates, through the external provider a voice call and SMS text message is sent to summon the attention of user-designated caregiver(s).

In addition to initiating alerts, the EpiMonitor app also continuously receives all the raw sensor data collected by the EmbracePlus. These data are analyzed by one of the EpiMonitor app software modules, EmpaDSP (paragraph 2.3.2), which computes the additional physiological parameters, such as activity during sleep and peripheral skin temperature.

The EpiMonitor App is also responsible for transmitting, over a cellular data plan or Wi-Fi connection the sensors' raw data, device information, and computed physiological parameters to the Empatica Cloud. On the Empatica Cloud, these data are stored, and made available to healthcare providers via a specific cloud-based software called Care Monitoring Portal.

Here's a summary of the acceptance criteria and study details for EpiMonitor, based on the provided FDA clearance letter:

Acceptance Criteria and Device Performance for EpiMonitor

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state "acceptance criteria" for PPA and FAR in a table format. Instead, it presents the device's performance for these metrics, implying that these results were deemed acceptable by the FDA for clearance. For the purpose of this response, I'm interpreting the "reported device performance" as the achieved PPA and FAR values and will frame the "acceptance criteria" as the expectation for these metrics to be within reasonable clinical utility.

2. Sample Size for the Test Set and Data Provenance:

• Epilepsy Monitoring Unit (EMU) Data (Retrospective Analysis):

  • Patients for PPA: 12 patients (6-21 years old) and 12 patients (>21 years old).
  • GTCS events for PPA: 19 GTCS events (6-21 years old) and 17 GTCS events (>21 years old).
  • Patients for FAR: 80 patients (6-21 years old) and 61 patients (>21 years old).
  • Data Provenance: Retrospective analysis of previously collected clinical data from patients observed in Epilepsy Monitoring Units. The document mentions data from "epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units" for the validation of the algorithm (EpiAlgo 2.1).

• Real-World Data (Longitudinal Analysis) for Low-Sensitivity Mode:

  • Patients for PPA/FAR: 601 patients (6-21 years old) and 843 patients (>21 years old).
  • GTCS events for PPA: 1157 GTCS events (6-21 years old) and 3625 GTCS events (>21 years old).
  • Observation days for FAR: 37594.2 days (6-21 years old) and 56389.1 days (>21 years old).
  • Data Provenance: Longitudinal analysis of real-world data, based on sensor data captured using the Embrace2 wearable device. This suggests the data was collected prospectively in a real-world setting, but its analysis for this specific submission was retrospective.

3. Number of Experts Used to Establish the Ground Truth and Qualifications:

• For the initial validation of EpiAlgo 2.1 (which supports the predicate device and is used in the subject device), the ground truth was "designed by a group of epileptologists at a top level 4 epilepsy center." The exact number of epileptologists and their specific years of experience are not provided. The method mentioned is "gold-standard video-Electroencephalogram (EEG) methodology."
• For the retrospective analyses presented, "adjudicated tonic-clonic seizure data" was used, implying expert review to establish the ground truth of GTCS events. The number and qualifications of the experts performing this adjudication for the analyses presented in Tables 1-4 are not explicitly stated.

4. Adjudication Method for the Test Set:

• The document implies clinical adjudication was performed to establish "adjudicated tonic-clonic seizure data" and the "gold-standard video-Electroencephalogram (EEG) methodology." However, it does not specify a particular adjudication method such as 2+1 or 3+1 for the test set data used in these retrospective analyses. It only mentions that the data was "adjudicated."

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study:

• No MRMC comparative effectiveness study was done.
• The document describes a standalone algorithm performance without human assistance for seizure detection.

6. Standalone (Algorithm Only) Performance:

• Yes, a standalone performance evaluation of the algorithm (EpiAlgo ver 2.1) was conducted. The PPA and FAR metrics presented (Tables 1-4) reflect the performance of the algorithm without human-in-the-loop assistance for seizure detection and alerting.

7. Type of Ground Truth Used:

• Expert Consensus / Clinical Diagnosis (Video-EEG): For the initial validation of EpiAlgo 2.1, the ground truth was established using "gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists." This indicates a high standard of clinical diagnosis and expert consensus.
• Adjudicated Data: For the retrospective analyses of EMU and real-world data, "adjudicated tonic-clonic seizure data" were used, implying expert review and decision-making on seizure events.

8. Sample Size for the Training Set:

• The document does not explicitly state the sample size for the training set of EpiAlgo ver 2.1. It mentions that EpiAlgo 2.1 was validated using data from epilepsy patients in EMUs, but this typically refers to validation/test sets, not specifically the training data.

9. How the Ground Truth for the Training Set Was Established:

• The method for establishing the ground truth for the training set is not detailed in this document. It only states that the EpiAlgo "has been validated through testing, using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units." This description primarily refers to the validation data, not the data used for initial training.

Ask a specific question about this device

The EpiWatch Monitoring System is a prescription, software-only mobile medical application intended for use with a compatible wrist-worn device as an adjunct to seizure monitoring of adults and children ages 5 and up in home or healthcare settings during periods of rest.

The EpiWatch Monitoring System continuously records, stores, displays, and transfers data from the compatible wristworn device's built-in physiological-based sensors to support review by healthcare professionals, and people with epilepsy (PWE) or at risk of epilepsy.

When the EpiWatch Monitoring System detects and logs physiological patterns associated with generalized tonic-clonic seizures (TCS), the EpiWatch Monitoring System application alerts the identified caregiver(s) to notify of detected possible seizure events.

The EpiWatch Monitoring System is a non-electroencephalogram (non-EEG) physiological signalbased seizure monitoring system. It is similar to other legally-marketated non-electroencephalogram (non-EEG) physiological signal-based seizure monitoring systems.

The EpiWatch Monitoring System consists of a software-only mobile medical application intended for use with a compatible wrist-worn device as an adjunct to seizure monitoring of adults and children ages 5 and up.

EpiWatch Monitoring System is a software platform composed of:

• A compatible wrist-worn device, (e.g. an Apple Watch)
• A mobile application running on smartphones called the "EpiWatch Monitoring System . App".

The wrist-worn device continuously collects raw data via specific sensors, and feeds consolidated physiological data to the EpiWatch app via APIs. Utilizing a proprietary algorithm, EpiWatch analyzes and assesses the physiological data and determines if there is suspected generalized tonicclonic seizure activity (TCS). The EpiWatch algorithm has been validated through testing, using the gold-standard video electroencephalogram (vEEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, of epilepsy patients experiencing TCSs in hospital Epilepsy Monitoring Units.

When a likely TCS is detected, the EpiWatch Monitoring System App communicates to the EpiWatch Cloud which initiates, through an external provider, a voice call, email and an SMS text message (notification parameters are set by the user at contact setup) to summon the attention of caregiver(s).

In addition to initiating alerts, the EpiWatch app also continuously receives all the sensor data collected by the wrist worn device. The EpiWatch App is also responsible for transmitting-over a cellular data plan or Wi-Fi connection-the API data, device information, and computed physiological parameters to the EpiWatch Cloud for further review and storage, which allows seizure reports and other features added to alerting. It also provides necessary information about the state of the system.

Here's a detailed breakdown of the EpiWatch Monitoring System's acceptance criteria and the study proving it, based on the provided document:

Acceptance Criteria and Device Performance

Note: The document only provides the reported device performance and does not explicitly state pre-defined acceptance criteria values for these metrics. However, these reported values are used as the basis for the substantial equivalence claim. The corrected PPA values are applied to account for extreme probability values and multiple seizures per patient, using the method of Saha et al. (2016).

Study Details

Based on the provided document, here's the information about the study that proves the device meets the acceptance criteria:

1. Sample size used for the test set and the data provenance:

   • Test Set Size: 242 subjects, with a total of 16,189 hours of study monitoring.
   • Data Provenance: The study was conducted in Epilepsy Monitoring Units (EMU) at six geographically diverse trial sites. The patients were in-patients experiencing TCSs. This suggests a prospective collection of data within a controlled clinical environment in the United States (implied by FDA submission and diverse trial sites without specifying other countries).
   • Exclusions: EMU patients with Lennox-Gastaut syndrome or Rett syndrome were excluded from analysis.

2. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

   • Number of Experts: Not explicitly stated, but the ground truth was "confirmed by the central panel" which was "designed by a group of epileptologists at a top level 4 epilepsy center."
   • Qualifications of Experts: Group of epileptologists at a top level 4 epilepsy center. General qualifications for "epileptologists" include specialized training and expertise in epilepsy. "Top level 4 epilepsy center" implies they are highly qualified and recognized in the field.

3. Adjudication method for the test set:

   • The ground truth was "confirmed by the central panel" and was based on "the gold-standard video electroencephalogram (vEEG) methodology." The specific adjudication method (e.g., 2+1, 3+1) used by the central panel is not explicitly stated.

4. If a multi reader multi case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance:

   • No, a MRMC comparative effectiveness study was not explicitly mentioned as being performed to compare human readers' performance with and without AI assistance. The study focuses on the standalone performance of the EpiWatch algorithm.

5. If a standalone (i.e. algorithm only without human-in-the loop performance) was done:

   • Yes, a standalone study was performed. The clinical study "evaluate[d] the performance of the EpiWatch algorithm for detecting TCSs," indicating it assessed the algorithm's performance independent of human-in-the-loop assistance for detection. The alerts are sent to caregivers, but the detection itself is algorithm-driven.

6. The type of ground truth used:

   • Expert Consensus and Pathology (vEEG): The ground truth was established by a "central panel" using "gold-standard video electroencephalogram (vEEG) methodology." vEEG is a clinical gold standard for diagnosing and classifying seizures, providing both electrophysiological and behavioral (video) evidence, which can be interpreted by experts to confirm seizure events. This falls under both expert consensus and a form of diagnostic/pathological evidence.

7. The sample size for the training set:

   • The document does not specify the sample size used for the training set. It only details the clinical study performed for performance evaluation (test set).

8. How the ground truth for the training set was established:

   • As the training set size is not provided, the method for establishing its ground truth is also not specified in this document. However, given the nature of the device, it would likely involve similar methodologies (e.g., vEEG interpreted by epileptologists), but this is not explicitly stated.

Ask a specific question about this device

EpiMonitor is a prescription only medical device system composed of a wearable device "EmbracePlus" and paired mobile software application "EpiMonitor" intended as an adjunct to seizure monitoring in adults and children aged 6 and up in a home environment or healthcare facilities. The device is worn on the wrist and senses Electrodermal Activity (EDA) and motion data to detect patterns that may be associated with either primary or secondary generalized tonic clonic clonic clonic seizures in patients with epilepsy or at risk of having epilepsy. When a seizure event is detected, the wearable device component of EpiMonitor sends a command to a paired mobile device where the EpiMonitor App is programmed to initiate an alert to a designated caregiver. The EpiMonitor app incorporates additional detection sensitivity modes, "high" for use during periods of rest or sleeping or "low" for use during periods of low-intensity activity, in order to reduce false alarm incidents.

EpiMonitor records, stores and transmits accelerometer. EDA, peripheral skin temperature and activity data for subsequent retrospective review by a trained healthcare professional via a cloud-based software.

The EpiMonitor system consists of a wearable device and mobile application:

• A wearable medical device called EmbracePlus,
• A mobile application running on smartphones called "EpiMonitor"

The EmbracePlus is worn on the user's wrist and continuously collects raw data via specific sensors, these data are continuously analyzed by an on-board algorithm (EpiAlgo 2.1), which assesses the physiological data and determines if the user may be undergoing a generalized tonic-clonic seizure (GTCS). The EpiAlgo has been validated through testing, using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units. When a likely GTCS is detected, EmbracePlus sends, via Bluetooth Low Energy, a message to the EpiMonitor app. The EpiMonitor app communicates to the Empatica Cloud which initiates, through the external provider a voice call and SMS text message is sent to summon the attention of userdesignated caregiver(s).

In addition to initiating alerts, the EpiMonitor app also continuously receives all the raw sensor data collected by the EmbracePlus. These data are analyzed by one of the EpiMonitor app software modules, EmpaDSP (paragraph 2.3.2), which computes the additional physiological parameters, such as activity during sleep and peripheral skin temperature.

The EpiMonitor App is also responsible for transmitting, over a cellular data plan or Wi-Fi connection the sensors' raw data, device information, and computed physiological parameters to the Empatica Cloud. On the Empatica Cloud, these data are stored, and made available to healthcare providers via a specific cloud-based software called Care Monitoring Portal.

Here's a breakdown of the acceptance criteria and study details for the EpiMonitor device, derived from the provided FDA 510(k) summary:

1. Table of Acceptance Criteria and Reported Device Performance

The acceptance criteria for the EpiMonitor device's Low-Sensitivity mode were evaluated based on Positive Percent Agreement (PPA) for seizure detection and False Alarm Rate (FAR) for both Epilepsy Monitoring Unit (EMU) data and real-world data.

Note: The document explicitly states "Analysis of performance for the Low-Sensitivity alerting mode in the EpiMonitor system demonstrated acceptable seizure detection accuracy and a reduced rate of false alerts." This implies that the reported performance met the sponsor's internal acceptance criteria for these metrics. Specific numerical thresholds for "acceptable" are not explicitly stated within the provided text.

2. Sample Sizes Used for the Test Set and Data Provenance

For Epilepsy Monitoring Unit (EMU) Data (Retrospective Analysis):

• Seizure Detection (PPA):
  • Patients: 24 (12 for age 6-21, 12 for age >21)
  • GTCS events: 36 (19 for age 6-21, 17 for age >21)
• False Alarm Rate (FAR):
  • Patients: 141 (80 for age 6-21, 61 for age >21)
  • Days of monitoring: 241.62 (88.94 for age 6-21, 152.68 for age >21)
• Data Provenance: The data was collected from patients observed in Epilepsy Monitoring Units. The exact geographic origin (country) is not specified, but the data was from "a top level 4 epilepsy center" (mentioned in device description for original EpiAlgo validation). This was a retrospective analysis of previously collected clinical data.

For Real-World Data (Longitudinal Analysis) - based on Embrace2 wearable device:

• Seizure Detection (PPA):
  • Patients: 1444 (601 for age 6-21, 843 for age >21)
  • GTCS events: 4782 (1157 for age 6-21, 3625 for age >21)
• False Alarm Rate (FAR):
  • Patients: 1444 (601 for age 6-21, 843 for age >21)
  • Days of monitoring: 93983.3 (37594.2 for age 6-21, 56389.1 for age >21)
• Data Provenance: "real-world data" captured using the Embrace2 wearable device, likely from home settings. The exact geographic origin (country) is not specified. This was a retrospective longitudinal analysis of real-world data.

3. Number of Experts Used to Establish Ground Truth for the Test Set and Qualifications

The document states that the EpiAlgo was validated "using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center". It also refers to "adjudicated tonic-clonic seizure data" for the EMU data. This implies that epileptologists were involved in establishing the ground truth.

• Number of experts: Not explicitly stated, but referred to as "a group of epileptologists."
• Qualifications of experts: "epileptologists at a top level 4 epilepsy center." No specific experience (e.g., 10 years of experience) is detailed.

4. Adjudication Method for the Test Set

The document mentions "adjudicated tonic-clonic seizure data" for the EMU study. However, the specific adjudication method (e.g., 2+1, 3+1) is not explicitly described in the provided text.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

• Was an MRMC study done? No.
• The study focuses on the performance of the algorithm itself (standalone and with different sensitivity modes), not on how human readers improve with or without AI assistance.
• The effect size of human readers improving with AI vs. without AI assistance is not applicable as this type of study was not performed.

6. Standalone (Algorithm Only) Performance Study

• Yes, a standalone study was done. The entire performance analysis for PPA and FAR presented in Tables 1-4 reflects the algorithm's performance (EpiAlgo ver 2.1) using the Low-Sensitivity mode, without human intervention in the detection process. The device detects an event, and the app initiates an alert; there's no mention of a human-in-the-loop directly influencing the detection sensitivity.

7. Type of Ground Truth Used

• Expert Consensus / Gold Standard (Video-EEG): The ground truth for seizure events was primarily established using gold-standard video-Electroencephalogram (EEG) methodology and "adjudicated tonic-clonic seizure data." This indicates expert consensus based on clinical and physiological evidence.

8. Sample Size for the Training Set

The provided text does not specify the sample size used for the training set of the EpiAlgo. It only describes the validation phases for the Low-Sensitivity mode.

9. How the Ground Truth for the Training Set Was Established

The provided text states: "The EpiAlgo has been validated through testing, using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units."

This implies that the training data's ground truth was established by epileptologists using video-EEG data from patients with generalized tonic-clonic seizures (GTCSs) in hospital Epilepsy Monitoring Units. This is consistent with clinical gold standards for seizure identification.

Ask a specific question about this device

The SPEAC® System is intended for use as an adjunct to seizure monitoring in adults in the home or healthcare facilities during periods of rest.

The non-EEG Physiological Signal Based Seizure Monitoring System continuously records and stores surface electromyographic (sEMG) data for subsequent review.

Trained healthcare professionals may use the electrophysiological sEMG data during a post-hoc review, with other contextual data, to characterize upper-extremity motor activity (UEMA) ipsilateral to the activity.

Audio data recorded during seizure monitoring may be available for review by a trained healthcare professional.

The device is to be used on the belly of the biceps muscle to analyze sEMG signals. When sEMG signal patterns associated with a unilateral, appendicular, tonic extension that could be associated with a GTC seizure are detected, the SPEAC System sends adjunctive alarms to alert caregivers.

Adjunctive alarms may be disabled by a physician order while continuing to record sEMG data for subsequent review.

The SPEAC System is a wireless, non-invasive, physiological, surface electromyography (sEMG) recording, monitoring, and alerting system to be used as an adjunct to seizure monitoring during periods of rest. The System continuously records and stores surface electromyographic (sEMG) data for subsequent review by a physician. Trained healthcare professionals may use the electrophysiological sEMG data, with other contextual data, to characterize seizures with upper-extremity motor activity ipsilateral to the device from other activity. SPEAC data gives healthcare professionals another diagnostic tool to characterize seizure events in a home or hospital setting.

The System continuously records and distributes sEMG data at 1,000 Hz (and audio around detected events) for post-hoc review by physicians (or other trained healthcare professionals) for the characterization of seizure events. A physician may perform post-hoc review of the SPEAC System data to characterize motor events that may be associated with seizures.

The seizure monitoring algorithm is able to send alarms to notify patients and caregivers when a pattern that may be associated with a generalized tonic-clonic (GTC) seizure is measured. Physicians may order the System with or without alarms and may order threshold adjustments to customize the level at which the System alarms.

Data collected by the System is uploaded to Brain Sentinel's secure remote storage, the Data Distribution System (DDS), and is remotely accessible for physician review. All patient data is cyber-secured within Microsoft Azure which is FedRAMP certified.

The SPEAC System remains the same with no alterations of any kind. The sEMG-based seizure monitoring algorithm is identical to the predicate SPEAC System. The purpose of this submission is to expand the indications for use based on clinical performance testing that was submitted to support a determination of substantial equivalence. When trained appropriately, clinicians may use the subject device to perform post-hoc analysis of the sEMG data from the device, with other contextual patient data, to characterize seizures with upper-extremity motor activity ipsilateral to the device from other activity.

The SPEAC System (K200276) is intended for use as an adjunct to seizure monitoring in adults. The system records and stores surface electromyographic (sEMG) data for subsequent review by trained healthcare professionals to characterize upper-extremity motor activity (UEMA) ipsilateral to the device. The system also sends alarms to alert caregivers when sEMG signal patterns associated with a unilateral, appendicular, tonic extension that could be associated with a Generalized Tonic-Clonic (GTC) seizure are detected.

Here's a breakdown of the acceptance criteria and the study details:

1. Table of Acceptance Criteria and Reported Device Performance

The document does not explicitly state formal "acceptance criteria" with specific numerical thresholds for the clinical study. Instead, the study evaluates the accuracy of expert reviewers in classifying seizure events using sEMG data. The reported device performance is presented as the accuracy of these expert reviewers.

2. Sample Size and Data Provenance

• Test Set Sample Size:

  • Individual assessments: 243 events (n=243)
  • Committee approach: 81 events (n=81)
  • Specific breakdown by event type (for committee):
    • Tonic-Clonic: 26
    • Simple Motor ES (Tonic & Clonic): 14
    • Complex Motor ES ("Other"): 22
    • All ES (Epileptic Seizures): 62
    • PNES, whole body involvement: 4
    • PNES, arm jerks/hand tremors only: 15
    • All PNES: 19

• Data Provenance: Prospective clinical trials conducted in an Epilepsy Monitoring Unit (EMU). The country of origin is not explicitly stated, but clinical trials subject to FDA review are typically conducted in the US or under internationally recognized standards.

3. Number of Experts and Qualifications

• Number of Experts: Three (3) sEMG reviewers participated in both studies.
• Qualifications of Experts: The document states "Physicians" for the sEMG review. For the ground truth, "Epileptologists" reviewed vEEG data, implying these are highly qualified medical professionals specializing in epilepsy. Specific years of experience are not mentioned for any of the reviewers.

4. Adjudication Method for the Test Set

• The study used a "committee" style approach, which is defined as majority rules (2/3) for combining the assessments of the three sEMG reviewers.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

• Yes, a form of MRMC study was done. The study involved multiple expert readers (three sEMG reviewers) evaluating multiple cases (seizure events).
• Effect Size of Human Readers with AI vs. without AI: This study's primary goal was not to measure the improvement of human readers with AI assistance. Instead, it evaluated the ability of clinicians to characterize events using sEMG data alone (from the device) and compared these characterizations to vEEG data and automated processing. The sEMG data from the SPEAC system is intended as an adjunct to monitoring, meaning human clinicians still interpret the data. Thus, an "AI vs. without AI assistance" effect size is not directly provided in the context of human reader improvement via an AI algorithm, but rather the utility of the sEMG data itself for expert interpretation.

6. Standalone (Algorithm Only) Performance

• Yes, a standalone component of performance was done. The clinical endpoints included comparison of the expert sEMG review to "automated event characterization". However, the results presented in the table specifically detail the "Expert Reviewer Accuracy" and indicate "No automated processing for seizure characterization was cleared in this 510(k)." This suggests that while automated processing was evaluated, its performance results are not provided here for clearance, and the focus for clearance is on the expert interpretation of the sEMG data.

7. Type of Ground Truth Used

• The ground truth for seizure characterization was established using Video-EEG (vEEG) data interpreted by epileptologists.

8. Sample Size for the Training Set

• The document does not specify a separate sample size for a training set. The study describes training physicians in interpreting sEMG data, but it refers to a prospective clinical trial where data was collected. It's unclear if a separate training data set for an algorithm was used, as the focus of the presented clinical performance is on human interpretation of sEMG data. The statement "The sEMG-based seizure monitoring algorithm is identical to the predicate SPEAC System" suggests the algorithm was already established, and this submission focused on expanding the indications for human interpretation of its output.

9. How Ground Truth for the Training Set Was Established

• If there was a training set for an algorithm, the document does not describe how its ground truth was established. For the training of physicians, the ground truth would implicitly be their consensus or established medical knowledge based on reference methods like vEEG, as discussed in the study.

Ask a specific question about this device

The SPEAC® System is indicated for use as an adjunct to seizure monitoring in adults in the home or healthcare facilities during periods of rest. The System records and stores surface electromyographic (sEMG) data for subsequent review by a trained healthcare professional.

The device is to be used on the belly of the biceps muscle to analyze sEMG signals that may be associated with generalized tonic-clonic (GTC) seizures. When sEMG signal patterns associated with a unilateral, appendicular, tonic extension that could be associated with a GTC seizure are detected, the SPEAC System sends adjunctive alarms to alert caregivers. Adjunctive alarms may be disabled by a physician order while continuing to record sEMG data for subsequent review.

The SPEAC® System, formerly known as the Brain Sentinel® Monitoring and Alerting System (Predicate), is a physiological, surface electromyography (sEMG) monitor with or without alarms that records and stores data for review by a physician for characterization of seizure events. The System records sEMG data at 1,000 Hz and distributes physiological data. Data can be analyzed with an algorithm using the default threshold or by a modified threshold ordered by the physician. The sEMG monitor is worn unilaterally on the belly of the patient's biceps and it analyzes for sEMG GTC seizures and provide local, remote, audible, and visual seizure alarms when a GTC Seizure pattern that may be associated with such seizures that are detected. The SPEAC System provides sEMG recordings and audio data to physicians (or other trained healthcare professionals) for post-hoc review so that they may quantify and qualify the types of seizure events that their patients experience. Every 24 hours, the sEMG monitor is removed from the patient and replaced with the second sEMG Monitor on the opposite arm of the patient. The sEMG that is removed after 24-hours is then attached to a Base Station. By connecting the sEMG Monitor to the Base Station, the monitor charges and the recorded data is downloaded to the Base Station. The recorded data is then automatically uploaded to Brain Sentinel's cloudbased storage unit, Data Distribution System (DDS), where they await review by a physician. All patient data is cyber-secured within Microsoft Azure which is FedRAMP certified.

Here's an analysis of the provided text to extract the acceptance criteria and study information:

Acceptance Criteria and Device Performance Study for the SPEAC® System

The information provided describes the Brain Sentinel, Inc. SPEAC® System, a non-EEG physiological signal-based seizure monitoring system. This 510(k) submission (K182180) emphasizes the substantial equivalence to its predicate device (DEN140033), also from Brain Sentinel.

The document focuses on demonstrating that the modified SPEAC® System remains substantially equivalent to the predicate device rather than presenting a new, comprehensive study with specific acceptance criteria directly tied to a new device performance study. Instead, it relies on the predicate's established performance and confirms that the changes to the subject device (SPEAC® System) do not negatively impact those established characteristics.

Therefore, the "acceptance criteria and reported device performance" as requested would primarily refer to the performance established for the predicate device, which is maintained by the subject device. The primary "study" that proves the device meets "acceptance criteria" here is a demonstration of substantial equivalence, relying heavily on the predicate's performance and verification that minor changes do not alter essential safety and effectiveness.

1. Table of Acceptance Criteria and Reported Device Performance

Given the nature of a 510(k) for substantial equivalence and the provided document, the "acceptance criteria" are implied to be the established performance characteristics of the predicate device, which the subject device is shown to maintain. The reported device performance is largely a re-affirmation of the predicate's performance, as the core seizure detection algorithm is identical.

2. Sample Size Used for the Test Set and Data Provenance

The document explicitly states: "The sEMG based seizure detection algorithm is identical to the predicate." It does not provide new clinical data or a new test set for the algorithm's performance. The performance metrics cited (e.g., alarm latency) appear to be derived from the studies that supported the predicate device (DEN140033). Therefore:

• Sample Size for Test Set: Not specified in this document for the algorithm's core performance, as it relies on the predicate's established performance. The "Performance Validation of SPEAC System" would likely have involved technical validation rather than a new clinical test set for algorithm accuracy.
• Data Provenance: Not specified in this document for the original algorithm validation. Given the company is U.S.-based (San Antonio, Texas), it is likely the original predicate studies were conducted in the U.S. The studies for the predicate device would have been prospective to demonstrate its initial effectiveness. This submission (K182180) focuses on equivalence rather than new prospective clinical data for algorithm performance.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications

This information is not provided in the current document. Since the algorithm is identical to the predicate, any expert review and ground truth establishment would have occurred during the predicate's development and regulatory clearance process (DEN140033).

4. Adjudication Method

This information is not provided in the current document. As with ground truth establishment, this would have been part of the predicate device's validation.

5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study

There is no indication of an MRMC comparative effectiveness study being performed for this 510(k) submission. The device is a "non-EEG physiological signal-based seizure monitoring system" with automated alarm capabilities, not a diagnostic imaging AI that assists human readers in interpretation. Its primary function is to detect specific sEMG patterns and alert caregivers, and to record data for subsequent review by a trained healthcare professional. The focus is on the device's ability to identify specific sEMG patterns, not on how it enhances human interpretation of complex data in a comparative setting.

6. Standalone (Algorithm Only) Performance Study

Yes, the algorithm only performance was established as part of the predicate clearance (DEN140033). The document states: "The sEMG based seizure detection algorithm is identical to the predicate." The performance claim for alarm latency ("The device provides an alert from -30.82 – 25.06 seconds, with an average of 5.34 seconds (SEM ± 2.86), following the onset of sEMG activity that may be associated with a GTC seizure.") directly reflects the standalone performance of this algorithm. This 510(k) reaffirms that this standalone performance is maintained.

7. Type of Ground Truth Used

The type of ground truth used for validating the sEMG patterns would typically be expert consensus or adjudicated clinical events, correlating the sEMG signals with observed generalized tonic-clonic (GTC) seizures. The document indicates that the system analyzes "sEMG signals that may be associated with generalized tonic-clonic (GTC) seizures" and identifies "sustained sEMG contraction patterns—during the tonic phase and early transition to the clonic phase—that are pathognomonic of GTC seizures." This strongly implies that the ground truth for detection was based on clinically confirmed GTC seizures, likely verified by neurologists or experts in epileptology, alongside simultaneous sEMG recordings.

8. Sample Size for the Training Set

The document does not provide the sample size for the training set. This information would have been part of the original development and validation of the algorithm for the predicate device.

9. How the Ground Truth for the Training Set Was Established

The document does not explicitly state how the ground truth for the training set was established. However, given the nature of the device and its indications, it is highly probable that the training data collected for the predicate device involved:

• Prospective collection of sEMG data from patients with confirmed epilepsy, particularly those prone to GTC seizures.
• Simultaneous video-EEG monitoring and direct clinical observation by trained medical staff to accurately identify and timestamp the onset and characteristics of GTC seizures.
• Annotation of sEMG recordings by experts (e.g., epileptologists, neurophysiologists) to delineate the specific sEMG patterns "pathognomonic of GTC seizures" as input for algorithm development.

This would involve a rigorous clinical process to ensure accurate correlation between the sEMG signals and the actual seizure events for both training and validation of the algorithm.

Ask a specific question about this device

The Embrace is a prescription only device that is indicated for use as an adjunct to seizure monitoring of adults and children age 6 and up in home or healthcare facilities during periods of rest. The device is worn on the wrist and senses Electrodermal Activity (EDA) and motion data to detect patterns that may be associated with generalized tonic clonic seizures in patients with epilepsy or at risk of having epilepsy. When a seizure event is detected, Embrace sends a command to a paired wireless device that is programmed to initiate an alert to a designated caregiver. The System records and stores data from Accelerometer, EDA, and Temperature sensors for subsequent review by a trained healthcare professional.

The Embrace is a wearable biosensor device that can capture, store, and wirelessly transmit sensor data via Bluetooth to a paired remote device. Embrace runs an on-board algorithm to continuously process sensor data and make a decision about whether the data might indicate a generalized tonic clonic seizure (GTCS). The algorithm has been validated on data, labeled using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 6 epilepsy centers, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units. When a likely GTCS is detected, the Embrace sends a message to the Alert smartphone application, which initiates calls and texts to summon the attention of designated caregivers. The device also enables patients to manually record seizure events, and provides contextual information related to activity and sleep.

The provided document describes the FDA 510(k) clearance for the Empatica Embrace device, a non-EEG physiological signal-based seizure monitoring system. The information below is extracted and organized to answer your request.

1. Table of Acceptance Criteria and Reported Device Performance

While the document doesn't explicitly state "acceptance criteria" in a separate section with specific numerical thresholds for clearance, the performance results presented indicate the observed efficacy. The implicit acceptance criteria would be for the device to perform well enough to demonstrate substantial equivalence to the predicate device and ensure safety and effectiveness.

2. Sample Size Used for the Test Set and Data Provenance

• Sample Size: 141 patients diagnosed with epilepsy.
  • 80 pediatric patients (ages 6-21 years, median: 13 years)
  • 61 adult patients (ages 22-63 years, median: 39 years)
• Data Provenance: The study was conducted in a hospital Epilepsy Monitoring Unit (EMU) setting. While the document doesn't explicitly state the country of origin, the listed submitter address is in Milan, Italy. The data is prospective as patients were enrolled and monitored during their EMU stay for the purpose of this study.
• Observed GTCS: 31 EMU patients experienced a total of 54 generalized tonic-clonic seizures (GTCSs). 110 EMU patients did not experience any seizure.
• Recorded Data: 141 patients provided overall 409 days (9,806 hours) of ACM (accelerometer) and EDA (electrodermal activity) measurements, with a median of 49.2 hours of data per patient.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications

• Number of Experts: A "panel of three readers" was used.
• Qualifications of Experts: The ground truth was established using "the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 6 epilepsy centers." This implies the "readers" were highly qualified medical professionals experienced in epilepsy diagnosis and seizure classification, likely epileptologists, given the standard of care in EMUs. Specific years of experience are not mentioned.

4. Adjudication Method for the Test Set

The document states "relative to a panel of three readers." While it doesn't explicitly detail the adjudication method (e.g., majority vote, independent assessment with reconciliation), the phrase "panel of three readers" suggests a consensus or majority agreement approach was used to establish the ground truth from the video-EEG data. It is not explicitly stated as 2+1 or 3+1, but rather that the ground truth was derived from the consensus of these three readers.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was Done, What Was the Effect Size of How Much Human Readers Improve with AI vs Without AI Assistance

• No MRMC comparative effectiveness study was described where human readers' performance with and without AI assistance was evaluated. This study focused on the standalone performance of the device's algorithm in detecting GTCS.

6. If a Standalone (i.e., Algorithm Only Without Human-in-the-Loop Performance) Was Done

• Yes, a standalone performance study was done. The reported PPA and FAR metrics represent the performance of the Embrace device's algorithm in detecting GTCS events based on its internal processing of EDA and motion data, without direct human intervention in the detection process. The device sends an alert when a likely GTCS is detected, which then initiates actions by caregivers. The study evaluates the accuracy of these automated detections.

7. The Type of Ground Truth Used

• The ground truth used was expert consensus based on gold-standard video-Electroencephalogram (EEG) methodology. This is considered a high-fidelity ground truth for seizure detection in an Epilepsy Monitoring Unit (EMU) setting. Every recorded seizure was classified as epileptic.

8. The Sample Size for the Training Set

The document states: "Embrace runs an on-board algorithm to continuously process sensor data and make a decision about whether the data might indicate a generalized tonic clonic seizure (GTCS). The algorithm has been validated on data, labeled using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 6 epilepsy centers, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units."

• The document describes the validation of the algorithm on data from specific sources but does not explicitly state the sample size used for the training set. The 141 patient dataset is specifically described as the "clinical testing" dataset, suggesting it was used for evaluation, not necessarily for training. It's common for validation data to be distinct from training data.

9. How the Ground Truth for the Training Set Was Established

• Similar to the validation set, the description implies that the data used for training (or at least development and refinement) of the algorithm was "labeled using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 6 epilepsy centers, from epilepsy patients experiencing GTCSs in hospital Epilepsy Monitoring Units." This indicates a high-quality, expert-driven ground truth establishment process for the data used to develop and refine the algorithm.

Ask a specific question about this device

The Embrace is a prescription only device that is indicated for use as an adjunct to seizure monitoring of adults in home or healthcare facilities during periods of rest. The device is worn on the wrist, and senses Electrodermal Activity (EDA) and motion data to detect patterns that may be associated with generalized tonic seizures in patients with epilepsy or at risk of having epilepsy. When a seizure event is detected, Embrace sends a command to a paired wireless device that is programmed to initiate an alert to a designated caregiver. The System records and stores data from Accelerometers, EDA, and Temperature for subsequent review by a trained healthcare professional.

The Embrace is a wearable biosensor device that can capture, store, and wirelessly transmit sensor data via Bluetooth to a paired remote device. Embrace runs an on-board algorithm to continuously process sensor data and make a decision about whether the data might indicate a generalized tonic clonic seizure (GTC). The algorithm has been validated on data, labeled using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, from epilepsy patients experiencing GTCs in hospital Epilepsy Monitoring Units. When a likely GTC is detected, the Embrace sends a message to the Alert smartphone application, which initiates calls and texts to summon the attention of designated caregivers.

Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:

1. A table of acceptance criteria and the reported device performance

Note on Acceptance Criteria: The document does not explicitly state numerical acceptance criteria for PPA and FAR. Instead, it reports the device's performance metrics and implicitly suggests that these values are deemed acceptable for marketing the device as "substantially equivalent" to a predicate device. The FDA typically relies on a comparison to predicate devices and clinical justification for novel devices.

2. Sample size used for the test set and the data provenance

• Sample Size for Test Set:
  • 135 patients diagnosed with epilepsy were enrolled.
  • Out of these, 22 patients experienced a total of 40 generalized tonic clonic seizures (GTCSs).
  • 113 patients did not experience any seizures during the study.
• Data Provenance:
  • Country of Origin: Not explicitly stated, but the study was conducted within an "Epilepsy Monitoring Unit (EMU)" in a "top level 4 epilepsy center," suggesting a clinical setting in a developed healthcare system (likely the US, given the FDA submission).
  • Retrospective or Prospective: The study enrolled patients and observed them for seizure events within the EMU, implying a prospective collection of data for the purpose of validating the device.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts

• Number of Experts: Three readers.
• Qualifications of Experts: They are referred to as a "panel of three readers." Although their specific qualifications (e.g., neurologists, epileptologists, years of experience) are not explicitly detailed in this document, the context of an "Epilepsy Monitoring Unit" and "gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists" strongly suggests they were highly qualified medical professionals specializing in epilepsy and EEG interpretation.

4. Adjudication method (e.g., 2+1, 3+1, none) for the test set

The document mentions "relative to a panel of three readers" for PPA calculation. This implies that the ground truth was established by consensus (or agreement) among these three readers. The specific adjudication method (e.g., if at least 2 out of 3 had to agree, or if all 3 had to agree, or if a tie-breaking fourth expert was used) is not explicitly stated. However, the use of a "panel" suggests a structured review process.

5. If a multi-reader multi-case (MRMC) comparative effectiveness study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance

• Was an MRMC comparative effectiveness study done? No. The study described focuses on the standalone performance of the Embrace device (i.e., the algorithm's ability to detect seizures) against a human-established ground truth. It does not evaluate how human readers' performance might change with or without the device's assistance.

6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done

• Yes, a standalone performance study was done. The performance metrics (PPA and FAR) directly reflect the algorithm's ability to detect GTC seizures based on sensor data. The device "runs an on-board algorithm to continuously process sensor data and make a decision about whether the data might indicate a generalized tonic clonic seizure (GTC)."

7. The type of ground truth used (expert consensus, pathology, outcomes data, etc.)

• Ground Truth Type: Expert consensus based on "gold-standard video-Electroencephalogram (EEG) methodology." This is a strong form of ground truth for epilepsy, as video-EEG is the primary method for diagnosing and classifying seizures in clinical practice. The data was "labeled using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists."

8. The sample size for the training set

The document states, "The algorithm has been validated on data, labeled using the gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center, from epilepsy patients experiencing GTCs in hospital Epilepsy Monitoring Units."

However, it does not explicitly state the sample size of the training set. It only describes the validation set (the 135 patients). It's common for regulatory submissions to omit detailed training set information, focusing instead on the independent validation (test) set.

9. How the ground truth for the training set was established

• The ground truth for the training set (implied, as the exact set is not detailed) was established using the "gold-standard video-Electroencephalogram (EEG) methodology designed by a group of epileptologists at a top level 4 epilepsy center." This indicates that the training data was meticulously labeled by highly qualified experts using the most reliable method for seizure detection.

Ask a specific question about this device

The Brain Sentinel Monitoring and Alerting System is indicated for use as an adjunct to seizure monitoring in adults in the home or healthcare facilities during periods of rest. The device is to be used on the belly of the biceps muscle to analyze surface electromyographs (sEMG) signals that may be associated with generalized tonic (GTC) seizures and to provide an alarm to alert caregivers of unilateral, appendicular, tonic extension that could be associated with a GTC seizure. The System records and stores sEMG data for subsequent review by a trained healthcare professional.

The Brain Sentinel Monitoring and Alerting System is a sEMG-based system for identifying sEMG activity that may be associated with generalized tonic-clonic seizures (GTCS). The device has two main components: the sEMG monitor and the base station. The sEMG monitor is worn on the patient's upper arm and monitors EMG activity in the arm via cutaneous electrodes connected to the sEMG monitor. Upon identification of sEMG activity, the monitor communicates wirelessly to the base station, which alerts a healthcare provider or caregiver in one or more ways (e.g., audible alarm, text message, e-mail, etc.).

Acceptance Criteria and Study to Prove Device Meets Acceptance Criteria for Brain Sentinel® Monitoring and Alerting System

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample Size Used for the Test Set and Data Provenance

The study utilized a prospective, multicenter, non-randomized design.

• Total Intent to Monitor (ITM) Cohort: 199 subjects.
• Properly Placed (PP) Cohort: 149 subjects (from the ITM cohort, after excluding improper placements).
• PP Cohort with GTC Seizure (Adults): 17 subjects (who experienced 21 GTC seizures). This adult PP subset is the primary test set for the acceptance criteria.
• Data Provenance: The data was collected from eleven (11) National Association of Epilepsy Centers (NAEC) Level IV Epilepsy Centers, making it prospective and originating from multiple US-based clinical sites.

3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications

• Number of Experts: Three independent neurologists.
• Qualifications of Experts: The document specifies "independent neurologists" and "independent epileptologists." While specific years of experience are not provided, "Level IV Epilepsy Centers" imply highly qualified specialists in epilepsy.

4. Adjudication Method for the Test Set

A majority rules approach was taken (2 out of 3 neurologists) to identify GTC seizures from the vEEG records. For each identified seizure, the time of bilateral, appendicular, tonic extension reported by each reviewer was averaged for comparison to the device's alert time.

For device placement assessment, three independent reviewers evaluated video images. If at least two out of the three independent reviewers classified the placement as proper, the data was included in the Properly Placed (PP) cohort.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done

No, an explicit Multi-Reader Multi-Case (MRMC) comparative effectiveness study focusing on human readers improving with AI vs. without AI assistance was not reported in this document. The study focused purely on the standalone performance of the device in identifying sEMG signals associated with GTC seizures. The device alerts caregivers, implying human involvement, but the study design presented does not evaluate the change in human performance with this assistance.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was done

Yes, a standalone performance study was done. The clinical study describes the device (algorithm) identifying sEMG activity and then comparing its alerts against the ground truth established by independent neurologists from vEEG records. The device's PPA and false alarm rates are calculated based on these automated alerts, indicating standalone algorithm performance. The device provides "an alarm to alert caregivers," but the performance metrics are calculated solely on the device's output.

7. The Type of Ground Truth Used

The ground truth used was expert consensus from vEEG records. Specifically, "Identification of GTC seizures was performed by three independent neurologists who reviewed the vEEG records of each subject's EMU stay to determine if and when generalized tonic clonic (GTC) seizures occurred."

8. The Sample Size for the Training Set

The document does not explicitly state the sample size for a training set. The study describes a "clinical study" used to evaluate the device's operating characteristics. It mentions that "Recorded sEMG data was post-processed at various threshold settings," suggesting that the algorithm for seizure detection might have been developed using unseen data or iteratively refined, but a distinct "training set" size as part of the regulatory submission is not detailed. The phrase "post processing sEMG data recorded from the clinical study to determine the operating characteristics" might imply that the clinical study data itself was used for evaluating operating points, rather than a separate, prior training phase with a distinct dataset.

9. How the Ground Truth for the Training Set Was Established

As the document does not explicitly describe a separate "training set" and its size, it consequently does not detail how the ground truth for such a training set was established. If the "clinical study" data was also used for initial algorithm development or refinement, the ground truth would likely have been established similarly to the test set: expert consensus from vEEG records. However, this is inferred, not explicitly stated for a training phase.

Ask a specific question about this device