Encevis

• K Number: K211452
• Device Name: Encevis
• Manufacturer: Austrian Institute of Technology GmbH
• Date Cleared: 2021-12-02 ( 206 days)
• Product Code: OMB , OLT , OMA
• Regulation Number: 882.1400
• Type: Traditional
• Panel: Neurology
• Reference & Predicate Devices: K132306 , K171720
• Predicate For: K240993

Intended Use

1. encevis is intended for the review, monitoring and analysis of EEG recordings made by electroencephalogram (EEG) devices using scalp electrodes and to aid neurologists in the assessment of EG. This device is intended to be used by qualified medical practitioners who will exercise professional judgment in using the information.

2. The seizure detection component of encevis is intended to mark previously acquired sections of adult (greater than or equal to 18 years) EEG recordings that may correspond to electrographic seizures, in order to assist qualified clinical practitioners in the assessment of EEG traces. EEG recordings should be obtained with a full scalp montage according to the standard 10/20-system.

3. The spike detection component of encevis is intended to mark previously acquired sections of the patients EEG recordings that may correspond to spikes, in order to assist qualified clinical practitioners in the assessment of EEG traces. The Spike Detection component is intended to be used in adult patients greater than or equal to 18 years. encevis Spike Detection performance has not been assessed for intracranial recordings.

4. encevis includes the calculation and display of a set of quantitative measures intended to monitor and analyze the EEG waveform. These include frequency bands, rhythmic and periodic patterns and burst suppression. These quantitative EEG measures should always be interpreted in conjunction with review of the original EEG waveforms.

5. The aEEG functionality included in encevis is intended to monitor the state of the brain.

6. encevis provides notifications on an on-screen display for seizure detection, quantitative EEG and aEEG that can be used when processing a record during acquisition. Delays of up to several minutes can occur between the beginning of a seizure, the occurrence of a spike or detection of quantitative EEG features and when the encevis notifications will be shown to a user. encevis notifications cannot be used as a substitute for real time monitoring of the underlying EEG by a trained expert.

7. encevis PureEEG (Artifact Reduction) is intended to reduce EMG and electrode artifacts in a standard 10-20 EEG recording. PureEEG does not remove the entifact signal, and is not effective for other types of artifacts. PureEEG may modify portions of waveforms representing cerebral activity. Waveforms must still be read by a qualified medical practitioner trained in recognizing artifact, and any interpretation or diagnosis must be made with reference to the original waveforms.

8. This device does not provide any diagnostic conclusion about the patient's condition to the user.

Device Description

encevis combines several modalities for viewing and analyzing EEG data in one integrated software package. Encevis consists of the following modalities: encevis EEG-Viewer, encevis artifact reduction (PureEEG), encevis seizure detection, encevis spike detection (EpiSpike), encevis rhythmic and periodic patterns, encevis aEEG, encevis frequency bands, and encevis Burst Suppression.

AI/ML Overview

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

Acceptance Criteria and Device Performance

For Seizure Detection:

Criteria Type	Acceptance Criteria (Non-inferiority margins)	Reported Device Performance (encevis vs. Predicate Persyst)
Positive Percentage Agreement (PPA)	PPA non-inferior to Predicate (margin 10%)	encevis PPA: 86.52% vs. Persyst PPA: 75.94% (non-inferior)
Negative Disagreement Rate (NDR)	NDR non-inferior to Predicate (margin 4 false detections/24h)	encevis NDR: 11.2 false detections/24h vs. Persyst NDR: 10.61 false detections/24h (non-inferior), lower is better

For Spike Detection:

Criteria Type	Acceptance Criteria (Non-inferiority margins)	Reported Device Performance (encevis vs. Predicate Persyst)
Positive Percentage Agreement (PPA)	PPA non-inferior to Predicate (margin 3%)	encevis PPA: 79.15% vs. Persyst PPA: 8.7% (non-inferior)
Negative Percentage Agreement (NPA)	NPA non-inferior to Predicate (margin 3%)	encevis NPA: 97.64% vs. Persyst NPA: 99.69% (non-inferior)
Positive Localization Percentage Agreement (PLPA)	PLPA non-inferior to Predicate (margin 3%)	encevis PLPA: 93.39% vs. Persyst PLPA: 93.97% (non-inferior)

For Artifact Reduction (PureEEG):

Criteria Type	Acceptance Criteria (Non-inferiority margins)	Reported Device Performance (encevis vs. Predicate Persyst)
Relative Suppression of Clean EEG (dB)	Non-inferior to Predicate (margin 1 dB)	encevis suppression lower than Persyst in 73/127 cases; 95% delta CI=[-0.07, -0.02], margin=0.01 (non-inferior)
Signal-to-Noise Ratio (SNR) after artifact removal (dB)	Non-inferior to Predicate (margin 1 dB)	encevis SNR higher than Persyst in 83/93 cases; 95% delta CI=[4.37,5.88], margin=0.01 (non-inferior)

For Rhythmic and Periodic Patterns (NeuroTrend):

Pattern Type	Sensitivity (%)	Specificity (%)
ANY	81.86 (79.9-83.8)	83.80 (83.1-84.5)
PD	69.73 (67.2-72.3)	95.89 (95.5-96.3)
ARA (RTA, RAA, RDA+S)	89.40 (84.2-94.6)	94.85 (94.5-95.3)
RDA	91.73 (86.4-97.1)	86.05 (85.4-86.7)
Localization (ACNS Main Term 1)	Cohen's Kappa: 0.51 (Moderate agreement) compared to expert consensus

For Burst Suppression:

Performance Metric	Value (%)
Sensitivity (SE)	87 (84.7-89.9)
Specificity (SP)	92 (91.4-92.9)
Positive Predictive Value (PPV)	61 (57.9-64.3)
Negative Predictive Value (NPV)	98 (97.7-98.5)

For aEEG (Amplitude-integrated EEG) and Frequency Bands:

These features were validated by comparing their output to the proposed method of Zhang and Ding (2013) and to the predicate device Persyst. The frequency response of aEEG was checked, and the outputs of both aEEG and frequency bands were found to be in good accordance with the reference and predicate, respectively. Specific numerical acceptance criteria were not explicitly stated beyond "good accordance" and "correct assignment," but the validation aimed to demonstrate consistency and accuracy.

---

Study Details

2. Sample sizes and data provenance for the test sets:

• Seizure Detection:
  • Sample Size: 55 subjects (scalp-EEG recordings). 50 patients with epilepsy, 5 subjects without epilepsy.
  • Data Provenance: Retrospective, from video-EEG monitoring in an epilepsy monitoring unit. Country of origin not specified, but the text mentions "different centers" for artifact reduction, rhythmic patterns, and burst suppression, implying multi-center data collection.
  • EEG Hours: A total of 1619 hours of EEG data were used. A maximum of 30 hours of continuous EEG data from each subject. For subjects without epilepsy, the first available 30 hours were included.
• Spike Detection:
  • Sample Size: 23 patients (scalp-EEG recordings). 18 subjects with spike events, 5 subjects without epilepsy.
  • Data Provenance: Retrospective, from video-EEG monitoring in an epilepsy monitoring unit. Country of origin not specified.
• Artifact Reduction (PureEEG):
  • Sample Size: 128 EEG data records from different patient groups (60 from epilepsy monitoring units, 65 from ICU patients).
  • Data Provenance: Retrospective. Includes 31 EEG segments from 31 subjects for seizure EEGs, 33 EEG segments from 6 subjects for spikes, and 65 EEG segments from 65 subjects for ICU patients. Country of origin not specified, but the text mentions "different centers".
• Rhythmic and Periodic Patterns (NeuroTrend):
  • Sample Size: 83 long-term EEGs from ICU patients, resulting in 11935 common annotation segments (each minute split into three 20-second segments).
  • Data Provenance: Prospective, from two different centers. Country of origin not specified.
• Burst Suppression:
  • Sample Size: 83 long-term EEGs from intensive care patients, resulting in 3978 valid annotation segments.
  • Data Provenance: Retrospective, from two different centers. Country of origin not specified.

3. Number of experts and qualifications for ground truth:

• Seizure Detection: 3 independent neurologists for blinded review.
• Spike Detection: 3 independent neurologists for blinded review.
• Artifact Reduction (PureEEG): 3 independent epileptologists or neurologists for blinded review.
• Rhythmic and Periodic Patterns (NeuroTrend): 2 clinical neurophysiologists, who were naive to these EEGs.
• Burst Suppression: 2 clinical EEG experts (reviewers).

Specific years of experience or detailed qualifications beyond "neurologist," "epileptologist," and "clinical neurophysiologist" were not provided in the text.

4. Adjudication method for the test set:

• Seizure Detection: An event was considered "true seizure" if the time interval of two out of three reviewers overlapped by at least 1 second. The seizure epoch was then defined as the overlapping time range of these two reviewers.
• Spike Detection: An event was considered as "true spike" only if the time interval of two out of three reviewers overlapped. (Localization information was based on the 3D-coordinates of the electrode next to the spike maximum, averaged over reviewers).
• Artifact Reduction (PureEEG): The text mentions that three EEG experts were engaged to identify "clean, pure EEG patterns and artifacts of different types," implying a consensus-based approach, but a specific adjudication rule (e.g., 2+1) is not explicitly detailed for this section.
• Rhythmic and Periodic Patterns (NeuroTrend): Annotations had to be consistent between both reviewers to be used in sensitivity and specificity measurement.
• Burst Suppression: Detection performance was analyzed for consensus annotations of the two reviewers, meaning only annotation segments where both reviewers showed the same decision about Burst Suppression pattern were used.

5. Multi-reader multi-case (MRMC) comparative effectiveness study:

• No, a full MRMC comparative effectiveness study was not explicitly described as per the traditional definition (i.e., measuring how human readers improve with AI assistance vs. without AI assistance).
• The studies primarily focused on standalone algorithm performance compared to a human-established ground truth and, in many cases, compared to a predicate device (Persyst).
• For Rhythmic and Periodic Patterns and Burst Suppression, the device's performance (sensitivity, specificity) was measured against an expert consensus, implying the device is intended to assist, but there's no direct evaluation of human performance with vs. without the device.
• Therefore, no effect size for human readers improving with AI vs. without AI assistance was reported.

6. Standalone (algorithm only without human-in-the-loop performance) study:

• Yes, standalone performance was done for all described features.
• For Seizure Detection, Spike Detection, and Artifact Reduction, the device's detection results were compared to the expert-established ground truth.
• For Rhythmic and Periodic Patterns and Burst Suppression, sensitivity and specificity were calculated based on the algorithm's detections against the expert consensus.
• For aEEG and Frequency Bands, the algorithm's output was compared to theoretical models and a predicate device.

7. Type of ground truth used:

• Expert Consensus:
  • Seizure Detection: Consensus annotations from three independent neurologists.
  • Spike Detection: Blinded review sessions from three neurologists.
  • Artifact Reduction (PureEEG): Annotations of clean EEG recordings and artifacts by three independent epileptologists or neurologists.
  • Rhythmic and Periodic Patterns (NeuroTrend): Annotations from two clinical neurophysiologists, requiring consistency between them.
  • Burst Suppression: Annotations from two clinical EEG experts, requiring consensus.
• Theoretical/Predicate Comparison:
  • aEEG: Compared to the proposed method of Zhang and Ding (2013) and compared to Persyst.
  • Frequency Bands: Tested against sinusoidal test data and manually selected representative EEG recordings from epilepsy/ICU patients, then compared to Persyst.

8. The sample size for the training set:

• The document does not explicitly state the sample size for the training set for any of the features. The provided information solely pertains to verification and validation activities, which typically use separate, unseen datasets for testing.

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

• Since the training set size and details are not provided, the method for establishing ground truth for any potential training data is also not described in the submission. The document focuses on the test data ground truth establishment.

Summary

{0}------------------------------------------------

December 2, 2021

Image /page/0/Picture/1 description: The image contains the logo of the U.S. Food and Drug Administration (FDA). On the left side of the logo is the Department of Health & Human Services logo. To the right of that is the FDA logo, which is a blue square with the letters "FDA" in white. To the right of the blue square is the text "U.S. FOOD & DRUG ADMINISTRATION" in blue.

Tilmann Kluge, Ph.D. Official Correspondent Austrian Institute of Technology GmbH Giefinggasse 4 1210 Vienna, Austria

Re: K211452

Trade/Device Name: Encevis Regulation Number: 21 CFR 882.1400 Regulation Name: Electroencephalograph Regulatory Class: Class II Product Code: OMB, OLT, OMA Dated: November 2, 2021 Received: November 4, 2021

Dear Tilmann Kluge:

We have reviewed your Section 510(k) premarket notification of intent to market the device referenced above and have determined the device is substantially equivalent (for the indications for use stated in the enclosure) to legally marketed predicate devices marketed in interstate commerce prior to May 28, 1976, the enactment date of the Medical Device Amendments, or to devices that have been reclassified in accordance with the provisions of the Federal Food, Drug, and Cosmetic Act (Act) that do not require approval of a premarket approval application (PMA). You may, therefore, market the device, subject to the general controls provisions of the Act. Although this letter refers to your product as a device, please be aware that some cleared products may instead be combination products. The 510(k) Premarket Notification Database located at https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm identifies combination product submissions. The general controls provisions of the Act include requirements for annual registration, listing of devices, good manufacturing practice, labeling, and prohibitions against misbranding and adulteration. Please note: CDRH does not evaluate information related to contract liability warranties. We remind you, however, that device labeling must be truthful and not misleading.

If your device is classified (see above) into either class II (Special Controls) or class III (PMA), it may be subject to additional controls. Existing major regulations affecting your device can be found in the Code of Federal Regulations, Title 21, Parts 800 to 898. In addition, FDA may publish further announcements concerning your device in the Federal Register.

Please be advised that FDA's issuance of a substantial equivalence determination does not mean that FDA has made a determination that your device complies with other requirements of the Act or any Federal statutes and regulations administered by other Federal agencies. You must comply with all the Act's requirements, including, but not limited to: registration and listing (21 CFR Part 807); labeling (21 CFR Part

{1}------------------------------------------------

801); medical device reporting of medical device-related adverse events) (21 CFR 803) for devices or postmarketing safety reporting (21 CFR 4, Subpart B) for combination products (see https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reportingcombination-products); good manufacturing practice requirements as set forth in the quality systems (QS) regulation (21 CFR Part 820) for devices or current good manufacturing practices (21 CFR 4, Subpart A) for combination products; and, if applicable, the electronic product radiation control provisions (Sections 531-542 of the Act); 21 CFR 1000-1050.

Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR Part 807.97). For questions regarding the reporting of adverse events under the MDR regulation (21 CFR Part 803), please go to https://www.fda.gov/medical-device-safety/medical-device-reportingmdr-how-report-medical-device-problems.

For comprehensive regulatory information about mediation-emitting products, including information about labeling regulations, please see Device Advice (https://www.fda.gov/medicaldevices/device-advice-comprehensive-regulatory-assistance) and CDRH Learn (https://www.fda.gov/training-and-continuing-education/cdrh-learn). Additionally, you may contact the Division of Industry and Consumer Education (DICE) to ask a question about a specific regulatory topic. See the DICE website (https://www.fda.gov/medical-device-advice-comprehensive-regulatoryassistance/contact-us-division-industry-and-consumer-education-dice) for more information or contact DICE by email ([e-mail]) or phone [phone] or 301-796-7100).

Sincerely,

for Jay Gupta Assistant Director DHT5A: Division of Neurosurgical, Neurointerventional and Neurodiagnostic Devices OHT5: Office of Neurological and Physical Medicine Devices Office of Product Evaluation and Quality Center for Devices and Radiological Health

Enclosure

{2}------------------------------------------------

Indications for Use

510(k) Number (if known) K211452

Device Name encevis

Indications for Use (Describe)

1. encevis is intended for the review, monitoring and analysis of EEG recordings made by electroencephalogram (EEG) devices using scalp electrodes and to aid neurologists in the assessment of EG. This device is intended to be used by qualified medical practitioners who will exercise professional judgment in using the information.

2. The seizure detection component of encevis is intended to mark previously acquired sections of adult (greater than or equal to 18 years) EEG recordings that may correspond to electrographic seizures, in order to assist qualified clinical practitioners in the assessment of EEG traces. EEG recordings should be obtained with a full scalp montage according to the standard 10/20-system.

3. The spike detection component of encevis is intended to mark previously acquired sections of the patients EEG recordings that may correspond to spikes, in order to assist qualified clinical practitioners in the assessment of EEG traces. The Spike Detection component is intended to be used in adult patients greater than or equal to 18 years. encevis Spike Detection performance has not been assessed for intracranial recordings.

4. encevis includes the calculation and display of a set of quantitative measures intended to monitor and analyze the EEG waveform. These include frequency bands, rhythmic and periodic patterns and burst suppression. These quantitative EEG measures should always be interpreted in conjunction with review of the original EEG waveforms.

5. The aEEG functionality included in encevis is intended to monitor the state of the brain.

6. encevis provides notifications on an on-screen display for seizure detection, quantitative EEG and aEEG that can be used when processing a record during acquisition. Delays of up to several minutes can occur between the beginning of a seizure, the occurrence of a spike or detection of quantitative EEG features and when the encevis notifications will be shown to a user. encevis notifications cannot be used as a substitute for real time monitoring of the underlying EEG by a trained expert.

7. encevis PureEEG (Artifact Reduction) is intended to reduce EMG and electrode artifacts in a standard 10-20 EEG recording. PureEEG does not remove the entifact signal, and is not effective for other types of artifacts. PureEEG may modify portions of waveforms representing cerebral activity. Waveforms must still be read by a qualified medical practitioner trained in recognizing artifact, and any interpretation or diagnosis must be made with reference to the original waveforms.

8. This device does not provide any diagnostic conclusion about the patient's condition to the user.

Type of Use (Select one or both, as applicable)

X Prescription Use (Part 21 CFR 801 Subpart D)

Over-The-Counter Use (21 CFR 801 Subpart C)

CONTINUE ON A SEPARATE PAGE IF NEEDED.

{3}------------------------------------------------

This section applies only to requirements of the Paperwork Reduction Act of 1995.

DO NOT SEND YOUR COMPLETED FORM TO THE PRA STAFF EMAIL ADDRESS BELOW.

The burden time for this collection of information is estimated to average 79 hours per response, including the time to review instructions, search existing data sources, gather and maintain the data needed and complete and review the collection of information. Send comments regarding this burden estimate or any other aspect of this information collection, including suggestions for reducing this burden, to:

Department of Health and Human Services Food and Drug Administration Office of Chief Information Officer Paperwork Reduction Act (PRA) Staff [e-mail]

"An agency may not conduct or sponsor, and a person is not required to respond to, a collection of information unless it displays a currently valid OMB number."

{4}------------------------------------------------

Image /page/4/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in a stylized, geometric font. The letters are in gray, and to the right of the letters is the text "AUSTRIAN INSTITUTE OF TECHNOLOGY" in a smaller, red font. The logo is simple and modern, and it is likely used to represent the organization's brand.

510k Summary encevis

{5}------------------------------------------------

Image /page/5/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a bold, sans-serif font, with the "A" stylized as a triangle. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The letters "AIT" are gray, while the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are maroon.

510k Summary encevis

1. Submission Sponsor and Application Correspondent

A. Submission Sponsor

AIT Austrian Institute of Technology GmbH Giefinggasse 4 1210 Vienna – Austria Phone: [phone] Fax: [phone] eMail: [e-mail]

2. Date Prepared

May 5th, 2021

Device Identification 3.

Trade/Proprietary Name:	encevis
Common Name:	Electroencephalograph
Classification Regulation:	21CFR882.1400
Product Code:	OMB, OLT, OMA
Class:	II
Panel:	Neurology

4. Legally Marketed Predicate Devices

Primary Predicate: Persyst 12 EEG Review and Analysis Software K132306 Additional Predicate: K171720 encevis 1.6

{6}------------------------------------------------

Image /page/6/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, sans-serif font. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The letters and words are in a dark gray color.

5. Device Description

encevis combines several modalities for viewing and analyzing EEG data in one integrated software package. Encevis consists of the following modalities:

encevis EEG-Viewer

The encevis EEG-viewer is intended for the review and the analysis of EEG-recordings that were recorded with an electroencephalography device using scalp electrodes. It shall aid the user in the examination of EEG-recordings. This includes frequency filtering of data, scaling of data in x and y direction and visualization in different montages. In addition, the encevis EEG-viewer can start several modules that automatically analyze the EEG and present the results in form of markers or in the form of modified EEG-curves. All included modules are intended for the user in the examination and monitoring of EEG-recordings.

encevis artifact reduction (PureEEG)

The artefact reduction encevis PureEEG is an analysis module that automatically recognizes and reduces artefacts in the EEG-data that come from EMG or electrode artefacts.

encevis seizure detection

The encevis seizure detection is a module for the automatic marking of areas in the EEG that could correspond to epileptic seizures with electrographic correlate, encevis seizure detection makes the results available to the user in form of marker list. The marker list is shown in the encevis EEG-viewer and in the EEG-trending user interface. This analysis can take place in parallel to the recording (ad-hoc) or after the recording finished (post-hoc).

encevis spike detection (EpiSpike)

The spike detection encevis EpiSpike is a module for the automatic marking of areas in the EEG that could correspond to pikes or spike-waves. A graphical user interface presents the reser. The user interface contains a time line per channel, a list of spike clusters that contain spikes and a list of spikes contained in a selected cluster. In addition, either the EEG or the averaged EEG 0.5 seconds before the spike maximum to 0.5 seconds after spike maximum for all spikes in a selected cluster is shown. This post-hoc analysis can take place in parallel to the recording or after the recording finished.

encevis rhythmic and periodic patterns

encevis rhythmic and periodic patterns is a feature for the analysis of EEG-recordings. It automatically detects EEG-patterns defined in the Standardized Critical Care EEG Terminology of the American Clinical Neurophysiology Society (Hirsch, L.J., et al., 2013. American Clinical Neurophysiology Society's Standardized Critical Care EEG Terminology: 2012 version. J. Clin. Neurophysiol. 30, 1–27) and graphically presents the results to the user. Additionally, it detects and visualizes rhythmic patterns with frequencies of up to 12Hz. It serves as a support during the examination of EEG-recordings in the ICU. This post-hoc analysis can take place in parallel to the recording or after the recording finished.

encevis aEEG

Encevis aEEG calculate and visualize a continuous measure that describes the EEG by showing the aEEG as defined in "Zhang, D., Dinq, H., 2013. Calculation of compact amplitude-integrated EEG

{7}------------------------------------------------

Image /page/7/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, geometric font in gray. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font, also in gray. The logo is simple and modern, reflecting the institute's focus on technology and innovation.

tracing and upper and lower margins using raw EEG data. Health (N. Y.) 05, 885–891. doi:10.4236/health.2013.55116".

encevis frequency bands

Encevis frequency bands calculate and visualize a continuous measure that describe the EEG by showing the frequency distribution of the EEG. The relative proportions of the four frequency bands Delta, Theta, Alpha, and Beta are shown. The intensity of the colors corresponds to the amplitudes in these four frequency bands.

encevis Burst Suppression

Encevis burst suppression feature calculates and visualizes suppression rate and suppression time in percent. It calculates periods of burst suppression in the EEG and marks them by vertical bars with red color. The definition of burst suppression patterns follows the guidelines of the American Clinical Neurophysiology Society ICU EEG Terminology (Hirsch, L.J., LaRoche, S.M., Gaspard, N., Gerard, E., Svoronos, A., Herman, S.T., Mani, R., Arif, H., Jette, N., Minazad, Y., Kerrigan, J.F., Vespa, P., Hantus, S., Claassen, J., Young, G.B., So, E., Kaplan, P.W., Nuwer, M.R., Fountain, N.B., Drislane, F.W., 2013. American Clinical Neurophysiology Society's Standardized Critical Care EEG Terminology: 2012 version. J. Clin. Neurophysiol. 30, 1-27. doi:10.1097/WNP.0b013e3182784729).

{8}------------------------------------------------

Image /page/8/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, geometric font. The letters are in gray, and to the right of the letters is the text "AUSTRIAN INSTITUTE OF TECHNOLOGY" in a smaller, burgundy font. The logo is simple and modern, and it effectively communicates the organization's name and focus.

Indication for Use Statement 6.

1. encevis is intended for the review, monitoring and analysis of EEG recordings made by electroencephalogram (EEG) devices using scalp electrodes and to aid neurologists in the assessment of EEG. This device is intended to be used by qualified medical practitioners who will exercise professional judgment in using the information.

2. The seizure detection component of encevis is intended to mark previously acquired sections of adult (greater than or equal to 18 years) EEG recordings that may correspond to electrographic seizures, in order to assist qualified clinical practitioners in the assessment of EEG traces. EEG recordings should be obtained with a full scalp montage according to the standard 10/20-system.

3. The spike detection component of encevis is intended to mark previously acquired sections of the patient's EEG recordings that may correspond to spikes, in order to assist qualified clinical practitioners in the assessment of EEG traces. The Spike Detection component is intended to be used in adult patients greater than or equal to 18 years. encevis Spike Detection performance has not been assessed for intracranial recordings.

4. encevis includes the calculation and display of a set of quantitative measures intended to monitor and analyze the EEG waveform. These include frequency bands, rhythmic and periodic patterns and burst suppression. These quantitative EEG measures should always be interpreted in conjunction with review of the original EEG waveforms.

5. The aEEG functionality included in encevis is intended to monitor the state of the brain.

6. encevis provides notifications on an on-screen display for seizure detection, quantitative EEG and aEEG that can be used when processing a record during acquisition. Delays of up to several minutes can occur between the beginning of a seizure, the occurrence of a spike or detection of quantitative EEG features and when the encevis notifications will be shown to a user. encevis notifications cannot be used as a substitute for real time monitoring of the underlying EEG by a trained expert.

7. encevis PureEEG (Artifact Reduction) is intended to reduce EMG and electrode artifacts in a standard 10-20 EEG recording. PureEEG does not remove the entifact signal, and is not effective for other types of artifacts. PureEG may modify portions of waveforms representing cerebral activity. Waveforms must still be read by a qualified medical practitioner trained in recognizing artifact, and any interpretation or diagnosis must be made with reference to the original waveforms.

8. This device does not provide any diagnostic conclusion about the patient's condition to the user.

{9}------------------------------------------------

Image /page/9/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, sans-serif font. The letters are in gray. To the right of the letters is the full name of the organization, "AUSTRIAN INSTITUTE OF TECHNOLOGY", in a smaller, sans-serif font. The words are stacked on top of each other, and the color is a dark red.

7. Substantial Equivalence Discussion

The following table compares the encevis to the predicate device with respect to intended use, technological characteristics and principles of operation, providing more detailed information regarding the basis for the determination of substantial equivalence.

	encevis(subject device)	Persyst 12(primary predicate)	Encevis 1.6(additional predicate)
510k Reference		K132306	K171720
Product Code	OMB	OMB	OMB
Additional Codes	OLT, OMA	OLT, OMA	OLT, OMA
Class	II	II	II
Regulation Number	21CFR882.1400	21CFR882.1400	21CFR882.1400
Regulation Name	Electroencephalograph	Electroencephalograph	Electroencephalograph
Manufacturer	AIT Austrian Institute ofTechnology GmbH	Persyst DevelopmentCorporation	AIT Austrian Institute ofTechnology GmbH
General Device Description	EEG Review and AnalysisSoftware	EEG Review and AnalysisSoftware	EEG Review and AnalysisSoftware
Shows EEG	YES	YES	YES
Has Artefact reduction	YES	YES	YES
Identifies seizures	YES	YES	YES
Identifies spikes	YES	YES	YES
Calculates quantitative EEGmeasures	YES	YES	YES
Calculated EEG measuresdisplayed	YES	YES	YES
Type of EEG-Analysis	Post-hoc analysis	Post-hoc analysis	Post-hoc analysis
Type of EEG	Scalp EEG	Scalp EEG	Scalp EEG
Population age	Adults (age > 18)	Adults (age >18);Spike detection for ages > 1month.	Adults (age > 18)
User	This device is intended to beused by qualified medicalpractitioners who will exerciseprofessional judgment in usingthe information.	This device is intended to beused by qualified medicalpractitioners who will exerciseprofessional judgment in usingthe information.	This device is intended to beused by qualified medicalpractitioners who will exerciseprofessional judgment in usingthe information.
Input Files	Display and calculation basedon EEG data recorded byexternal EEG systems. Theyare either read from the EEG-file provided by the EEGsystem or can be send toencevis using the interfaceprovided by AIT(AITInterfaceDLL)	Display and calculation isbased on EEG data recordedby external EEG systems.They are read from the EEG-file provided by the EEGsystem	Display and calculation basedon EEG data recorded byexternal EEG systems. Theyare either read from the EEG-file provided by the EEGsystem or can be send toencevis using the interfaceprovided by AIT(AITInterfaceDLL)
Compliance	No standard data formatavailable in the industry	No standard data formatavailable in the industry	No standard data formatavailable in the industry
Output Files	Results are stored in adatabase and/or is send overthe interface AlTInterfaceDLLto an external EEG system.User output is given bygraphical user interfacos	Results are stored inadditional files in the filesystem placed in the samefolder as the EEG file. Useroutput is given by graphicaluser interfaces	Results are stored in adatabase and/or is send overthe interface AITInterfaceDLLto an external EEG system.User output is given bygraphical usor interfaces
Compatible Equipment andSoftware	Encevis can read and processEEG data from several EEGvendors. A list of compatibleEEG systems can be found onhttp://www.encevis.com	Persys can read and processEEG data from several EEGvendors. A list of compatibleEEG systems can be found onhttp://www.persyst.com/support/supported-formats/	Encevis can read and processEEG data from several EEGvendors. A list of compatibleEEG systems can be found onhttp://www.encevis.com

{10}------------------------------------------------

Image /page/10/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, sans-serif font, with the "A" represented by a triangle. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The logo is simple and modern, and the use of the triangle in the "A" gives it a sense of forward movement and innovation.

Table 1: Comparison between encevis and the predicate device

Non-Clinical Performance Data 8.

Software verification and validation testing was conducted and documentation provided as recommended by the FDA Guidance for Industry and FDA Staff, Guidance for the Content of Software Contained in Medical Devices. Traceability has been documented between all system specifications to validation test protocols. Verification and validation testing includes:

• 1. Code inspections
• 2. Unit level testing
• 3. Integration level testing
• 4. System level testing

In addition, tests according to "IEC 62366-1:2015, Medical Devices Part 1-Application of usability engineering to medical devices" have been performed.

The software for this device is determined as a "moderate" level of concern because a failure or latent flaw could lead to a minor injury to the patient through incorrect information or through the action of the care provider.

Verification and validation activities established the safety and performance characteristics of the subject device with respect to the predicate device. The following performance data have been provided in support of the substantial equivalence determination.

FeatureTest	əmə	olise personalisme	ən bir sənələri və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşı və qarşılır. Bu və bir və qarşı	in fəsiləsinə cinsinə aid bitki növü. İstinadlar Respublikası Respublikası Respublikası Respublikası Respublikası Say Almaniya Alanmada Alanmada Alanma	uləşməd olmuşdur. Bu başları Şaman Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şuraq Şura	בה בינוש במש	spueq (suənbə)	ərin əsasən qarşı qarşı qarşı qarşı qarşı və qarşı qarşı qarşı qarşı qarşı qarşı və başların qarşı və başların qarşı qarşı və başların qarşı qarşılır. Bu və başların qarşı q
Direct comparison with predicatedevice		×	×	×		×		×
Bench test on large amount ofEEG data		×	×	×	×	×	×	×
Software test (System,Integration, Unit)	×	×	×	×	×	×	×	×

Table 2: Type of performance test per feature

For bench tests, detection results of the modules were compared to annotations set by clinical EEG experts using large amount of EEG data from different centers. Where possible, the results of encevis

{11}------------------------------------------------

Image /page/11/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in a stylized, geometric font in gray. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font, with "AUSTRIAN INSTITUTE" on top of "OF TECHNOLOGY."

were directly compared with the results of the predicate device. Suitable statistical measures like sensitivity and specificity were calculated.

The encevis (stand-alone software) meets all the stated requirements for overall design, performance, biocompatibility and electrical safety and passed all the testing noted above.

9. Clinical Performance Data

Seizure detection performance testing:

For performance evaluation of the encevis seizure detection device we measured positive percentage agreement (detection sensitivity based on the reference standard) and negative disagreement rate (false detections per 24 hours based on reference standard) by comparing seizure detections to consensus annotations from three independent reviewers. Second, to define the acceptable performance level of the encevis seizure detection device we also measured positive percentage agreement and neqative disagreement rate of the predicate device Persyst using the same study population and the same gold standard annotations. A statistical test is then used to show that the encevis seizure detection performance is non-inferior to the performance of the predicate device.

Study population

We included scalp-EEG recordings of 55 subjects that underwent video-EEG monitoring in an epilepsy monitoring unit for the purpose of differential diagnosis or pre-surgical evaluation. All patients where 18 years of age or older. 50 patients where included that showed seizure events during recording and were diagnosed of having epilepsy. Further, we included the five subjects that were diagnosed of not having epilepsy (Subject-ID 30-34). No further selection of subjects was made.

Reference Standard

To define the reference standard, a total of 1619 hours of EEG from these 55 subjects were presented to three independent neurologist for blinded review. The goal of the review sessions was to identify the start and end times of epileptic seizures to define "true seizure" epochs for later performance evaluation of the automatic seizure detection algorithm. The 1619 hours of EEG consisted of a maximum of 30 hours of continuous EEG data from each subject. For subjects without epilepsy the first available 30 hours of recording were included. The EEG experts were asked to mark the time positions of the seizure onset and end. An event was considered as "true seizure" only if the time interval of two out of three reviewers overlapped by at least 1 second. A seizure epoch was then defined as the overlapping time range of two reviewers.

The following tables show the number of EEG hours and seizures per subject.

{12}------------------------------------------------

Image /page/12/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in a stylized, geometric font. The letters are in a gray color. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, red font. The logo is simple and modern.

Image /page/12/Figure/3 description: The image contains two bar charts. The chart on the left is titled "Number of EEG hours per subject" and shows the length of EEG test data in hours for different subjects, with most subjects having around 30 hours of data. The chart on the right is titled "Number of seizures per patient" and shows the number of seizures for different subjects, with subject 11 having the highest number of seizures at 29. Both charts have the subject numbers on the y-axis, ranging from 1 to 55.

Figure 1: Number of EEG hours and seizures per subject

{13}------------------------------------------------

Image /page/13/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, sans-serif font, with the "A" formed by two diagonal lines. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The letters "AIT" are gray, while the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are maroon.

Inter-rater agreement

Inter-rater agreement was measured between any pair of reviewers using Cohen's kappa (x). The agreement values were к=0.69 (95% Cl=[0.62, 0.77]) between reviewer 2, к=0.76 (95% Cl=[0.69, 0.83]) between reviewer 1 and reviewer 3, and K=0.79 (95% Cl=[0.72, 0.85]) between reviewer 2 and reviewer 3. The average agreement resulted in K=0.74. According to the qualitative classification of Landis and Koch (Landis JR, Koch GG, Biometrics. 1977 Mar;33(1):159-74) the average к value of 0.74 can be interpreted as substantial agreement.

Detection Performance

To define positive percentage agreement (PPA) and negative disagreement rate (NDR, given as false detections in 24 hours) for each patient the seizure epochs defined by consensus annotations of two out of three reviewers were compared to automatically calculated seizure time points of the encevis seizure detection device and the predicate device Persyst. The encevis seizure detection device results in a single time point for each detection that is used in this validation. The predicate device Persyst was used with default settings (perception score = 0.5) and the given start time point was used in this validation. The logical variables true positive (FP), false positive (FP), and false negative (FN) are defined as follows: seizure epochs are counted as TP if at least one detection occurred within the epoch time range. Detections outside of seizure epochs were defined as false positives (FP). Seizure epochs without a matching detection were defined as false negative (FN).

Results

The average positive percentage agreement of the subjects with at least one "true seizure" event resulted in 86.52% (95% Cl=78.54, 94.49)) for encevis seizure detection and in 75.94% (95% Cl=65.5, 86.4]) for the predicate device Persyst.

The average negative disagreement rate (NDR) was 11.2 false detections in 24 hours (95% Cl=[7.04.15.3]) for the encevis seizure detection and 10.61 false detections in 24 hours (95% Cl=[6.8, 14.5]) for predicate device Persyst. The following two figures show the average results including confidence intervals.

{14}------------------------------------------------

Image /page/14/Figure/0 description: This image is a graph comparing the average positive percentage agreement between encevis and Persyst, including 95% confidence intervals. The x-axis represents the positive percentage agreement, ranging from 0 to 100. The graph shows that encevis has a higher average positive percentage agreement than Persyst. The average positive percentage agreement for encevis is around 85%, while the average positive percentage agreement for Persyst is around 75%.

Figure 2: Average and the confidence interval of the positive agreement performance between encevis and Persyst.

Image /page/14/Figure/2 description: The image is a plot comparing the average negative disagreement rate and 95% confidence intervals for two systems, encevis and Persyst. The x-axis represents the average negative disagreement rate, measured in false detections per 24 hours, ranging from 0 to 16. For encevis, the average negative disagreement rate is around 11, with a confidence interval spanning approximately from 7 to 15. For Persyst, the average negative disagreement rate is around 10, with a confidence interval spanning approximately from 7 to 14.

Figure 3: Average and the confidence interval of the negative disagreement performance between encevis and Persyst. Low numbers are better.

A Two One-Sided Test (TOST) procedure for paired samples (Walker E, Nowacki AS, J Gen Intern Med. 2011 Feb;26(2):192-6) was used to test the non-inferiority of the encevis seizure detection device to the predicate device. For statistical comparison, a type I error of 0.05 and non-inferiority margins of 10% for positive percentage agreement (PPA) and 4 for negative disagreement rate (NDR) were used.

{15}------------------------------------------------

Image /page/15/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in a stylized, geometric font. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The logo is simple and modern, and the colors are muted.

The two performance parameters PPA and NDR were tested independently to measure the noninferiority of both device parameters.

Both device parameters PPA and NDR of the encevis seizure detection are found to be non-inferior to the parameters of predicate device Persyst.

encevis spike detection performance testing

The clinical truth was determined based on the results of blinded review sessions from three neurologists. The "true spike" events (clinical truth) were then compared to automatically calculated spike time points of the encevis spike detection device to define true positives (TP), false positives (FP), false negatives (FN), and true negatives (TN) for each patient. With these values the positive percentage agreement (PPA) and negative percentage agreement (NPA) for each patient are calculated. In addition, "true spike" events were compared to the automatic detections of Persyst resulting in PPA and NPA values for the predicate device. Furthermore, the localization performance of both systems encevis spike detection and Persyst was evaluated based on the localization information given by the detection systems and the spatial information provided by the reviewer (clinical truth). We define a positive localization percentage agreement (PLPA) which is calculated for each patient.

Study population

To prove the validity of the spike detection system, encevis spike detection was tested with the EEG of 23 patients. For clinical validation, we included scalp-EEG recordings of 23 subjects that underwent video-EEG monitoring in an epilepsy monitoring unit for differential diagnosis or pre-surgical evaluation. 18 subjects of 18 years of age or older that showed spike events during recording based on initial clinical information where included. In addition, five subjects of 18 years of age or older that were diagnosed of not having epilepsy were included (Subject-ID 9-13). No further selection of subjects were made.

The statistical parameters PPA, NPA and PLPA were used in a two one-sided test (TOST, (Walker E et. al.) using paired samples in order to show the non-inferiority of encevis spike detection device compared to the predicate of Persyst.

Reference standard

To define the clinical truth the EEG from all subjects were presented to three independent Neurologists for blinded review. The goal of the review sessions was to identify all "true focal spikes" for later performance evaluation of the automatic spike detection algorithm. The EEG experts were asked to mark the time positions at the beginning and the end of the spike. Furthermore, the reviewers were asked to specify the electrode which is next to the spike maximum (phase reversal).

An event was considered as "true spike" only if the time interval of two out of three reviewers overlapped. For the determination of the localization performance. the 3D-coordinates of the electrode which is next to the spike maximum averaged over reviewers was used. The determined average position is considered as the clinical truth with respect to the localization and is used to evaluate the localization performance of encevis spike detection and the predicate Persyst.

Performance evaluation

Data of all 23 subjects was processed with encevis spike detection. In order to compare the obtained results of encevis spike detection with the predicate Persyst, the same data was processed with the spike detector of Persyst 12. The detection systems were evaluated by means of suitable performance measures like positive percentage agreement (PPA) and negative percentage agreement (NPA). For measuring the localization performance, we defined a positive localization percentage agreement

{16}------------------------------------------------

Image /page/16/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, geometric font in gray. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font in red. The logo is simple and modern, and the colors are muted and professional.

(PLPA). The basis for the performance evaluation are the annotations of the EEG experts which were placed at the onset and the end of the spike. Comparison between the time instances of the annotations and the time instances of automatic detections allows assessing the performance. The detection resolution of both systems, encevis spike detection and the Persyst spike detection was one microsecond.

Results of the performance measures

The average positive percentage agreement of the 15 subjects with at least one "true spike" event resulted in 79.15% (95% Cl=[67.7-90.6]) for encevis spike detection and in 8.7% (95% Cl=[4.4-13.0]) for the predicate device Persyst.

The average negative percentage agreement of all 23 subjects was 97.64 (95% Cl=[96.6.-98.6]) for the encevis spike detection and 99.69 (95% Cl=[99.4-99.9]) for predicate device Persyst.

The average positive localization percentage agreement of the 12 subjects with at least one "true positive" event was 93,39 (95% Cl=[84.8.-102.0]) for the encevis spike detection and 93.97 (95% Cl=[83.6-104.3]) for predicate device Persyst.

In the following figures the averages and the confidence intervals for the PPA, NPA and PLPA are visualized.

Image /page/16/Figure/9 description: The image is a plot comparing the average positive percentage agreement between two entities, Persyst and encevis, including 95% confidence intervals. Persyst has an average positive percentage agreement of around 10%, while encevis has an average positive percentage agreement of around 80%. The x-axis represents the average positive percentage agreement in percentage, ranging from 0 to 100. The title of the plot is "Average positive percentage agreement (%) including 95% confidence intervals".

Figure 4: This figure compares the average and the confidence interval of the positive percentage agreement performance between encevis and Persyst.

{17}------------------------------------------------

Image /page/17/Figure/0 description: The image shows a plot comparing the average negative percentage agreement between "encevis" and "Persyst". The x-axis represents the average negative percentage agreement in percent, ranging from 50 to 100. Both "encevis" and "Persyst" have an average negative percentage agreement close to 100, with "encevis" having a confidence interval shown.

Figure 5: This figure compares the average and the confidence interval of the negative percentage agreement performance between encevis and Persyst.

Image /page/17/Figure/2 description: The image is a bar chart that shows the average positive localization percentage agreement, including 95% confidence intervals. The x-axis shows the average positive localization percentage agreement in percentage, and the y-axis shows the names of two systems, "encevis" and "Persyst". The average positive localization percentage agreement for "encevis" is around 92%, and the average positive localization percentage agreement for "Persyst" is around 93%.

Figure 6: This figure compares the average and the confidence interval of the positive localization percentage agreement performance between encevis and Persyst.

A Two One-Sided Test (TOST) for paired samples (Walker et al) was used to test the non-inferiority of the encevis spike detection device to the predicate device. For statistical comparison, a type I error of 0.05 and non-inferiority margins of 3% for positive percentage agreement (PPA), the negative percentage agreement (NPA) and the positive localization percentage agreement (PLPA) are used. The three performance measures PPA, NPA and PLPA were tested independently to measure the noninferiority of all device parameters separately.

{18}------------------------------------------------

Image /page/18/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a gray, sans-serif font. To the right of the letters is the text "AUSTRIAN INSTITUTE OF TECHNOLOGY" in a smaller, red, sans-serif font. The logo is simple and modern, and it is likely used on the organization's website, marketing materials, and other official documents.

encevis artifact reduction performance testing:

The quality of an artifact removal algorithm is determined by two aspects.

• 1. The method should not significantly modify true, clean EEG pattern that are not disturbed by artifacts. To quantify the performance of the algorithms with regard to this aspect, changes in clean EEG patterns due to the algorithms are evaluated.
• 2. The method should suppress artifacts that are superimposed on the true EEG as far as possible, revealing the underlying, pure EEG patterns. To quantify the algorithms to remove artifacts, signal-to-noise ratios will be measured before and after artifact removal.

For these measurements we need clean, pure EEG patterns and artifacts of different types. In order to identify these patterns, three EEG experts Neurologists are engaged as independent reviewers.

Validation data

For the validation study, 128 EEG data records from different patient groups are used, covering all intended use populations of encevis, i.e., adult patients in epilepsy monitoring and in critical care. Each record consisted of 10 seconds of data to be evaluated. The datasets include 60 patients from epilepsy monitoring units and 65 from ICU patients. These data were selected as follows:

Epilepsy monitoring – seizure EEGs: We include 31 EEG segments from 31 subjects of 18 years of age or older that underwent video-EEG monitoring in an epilepsy monitoring unit for the purpose of differential diagnosis or pre-surgical evaluation and that showed seizure events during recording and were diagnosed of having epilepsy.

Epilepsy monitoring - spikes: We include 33 EEG segments from 6 subjects of 18 years of age or older that underwent video-EEG monitoring in an epilepsy monitoring unit for the purpose of differential diagnosis or pre-surgical evaluation that showed spikes during recording.

Intensive care unit: We include 65 EEG segments from 65 subjects of 18 years of age or older that have been admitted to an intensive care unit due to severe neurological disorders (cerebral hypoxia, cerebral ischemia, cerebral hemorrhage of different genesis, cerebral tumors, status epilepticus, infections, toxidromes, encephalopathies of different genesis, cerebral malformations and craniocerebral traumas) on a systemic or localized basis. The random selection includes 9 segments with seizures, 10 segments with rhythmic activity, 11 segments with periodic discharges, 17 segments with burst-suppression and 18 segments without any pattern.

Expert review

For this validation study we need annotations of clean EEG recordings without any artifacts, and moreover annotations of artifacts that can be superimposed to the clean recordings. We engage three independent epileptologists or neurologists for blinded review of the EEG data from EMU and ICU.

Statistical testing

A Two One-Sided Test (TOST) procedure for paired samples (Walker E, Nowacki AS, J Gen Intern Med. 2011 Feb;26(2):192-6) is used to test the non-inferiority of the encevis artifact reduction compared to the predicate device. For statistical comparison, a type I error of 0.05 and non-inferiority margins of 1dB.

The hypothesis to test non-inferiority of the relative suppression of true EEG in dB is defined as:

• . H0: The relative suppression of true EEG in dB of the encevis artifact removal is higher than the suppression of true EEG in dB of the predicate device.

{19}------------------------------------------------

Image /page/19/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in a gray sans-serif font. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, dark red sans-serif font. The logo is simple and modern, and the colors are muted and professional.

• . H1: The relative suppression of true EEG in dB of the encevis artifact removal is lower than or equal to the suppression of true EEG in dB of the predicate device.
  The hypothesis to test the signal-to-noise ratio after artifact removal is defined as:

• . H0: The signal-to-noise ratio after artifact removal by encevis is lower than the signal-to-noise ratio after artifact removal by of the predicate device.

• . H1: The signal-to-noise ratio after artifact removal by encevis is higher than or equal to the signal-to-noise ratio after artifact removal by of the predicate device.

The results of the evaluation of relative suppression of clean EEG are summarized in the following table in %. This number means, that the variance of the clean EEG activity has been suppressed by this relative value, i.e., low values are desired. Due to technical reasons, only 127 out of 131 test cases could be evaluated: in the remaining 4 cases, Persyst produced zero lines in all channels.

Image /page/19/Figure/8 description: The image is a horizontal bar chart that shows the relative suppression of clean EEG in percentage. The x-axis represents the percentage of relative suppression, ranging from 0.00% to 80.00%. The y-axis represents different data points, labeled with numbers from 1 to 129. The chart compares two datasets, "encevis" and "Persyst", with each data point having two bars representing the relative suppression for each dataset.

Figure 7: Relative suppression of clean EEG by encevis and Persyst

{20}------------------------------------------------

Image /page/20/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, geometric font. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The logo is simple and modern, and the colors are muted.

The results of the evaluation SNR prior and post artifact removal are summarized in the following table in dB. This numbers show the signal-to-noise ratio (noise=artifacts) that has been achieved after artifact removal, i.e., high values are desired. Eleven out of 104 test cases could not be evaluated, since the artifacts in these cases were on channels, where the initial EEG was not undistorted according to reviewers.

Image /page/20/Figure/4 description: The image is a bar graph titled "SNR prior and post artifact removal [dB]" that compares the initial SNR to the SNR after artifact removal using two different methods: encevis and Persyst. The x-axis represents the SNR in decibels, ranging from -60 to 30. The y-axis lists numbers from 1 to 104, with some numbers missing. The graph shows the SNR values for each method at different points, with blue bars representing the initial SNR, red bars representing encevis, and green bars representing Persyst.

Figure 8: SNR prior and post artifact removal by encevis and Persyst

{21}------------------------------------------------

Image /page/21/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, geometric font. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, sans-serif font. The logo is simple and modern, and the colors are muted.

Results of Statistical testing

The results of the Two One-Sided Test for relative Suppression of clean EEG (Test-Control) are (95% delta CI=[-0.07, -0.02], margin = 0.01):

$-0.02 < 0.01 -> H1$

The results of the Two One-Sided Test for signal-to-noise ratios after artifact removal are (95% delta Cl=[4.37,5.88], margin = 0.01):

$4.37 > -0.01 -> H1$

Both device parameters, "relative Suppression of clean EEG" and "signal-to-noise ratios after artifact removal" of the encevis artifact reduction are therefore non-inferior to the parameters of predicate device Persyst.

In the statistical evaluation of both device parameters, "relative Suppression of clean EEG" and "signalto-noise ratios after artifact removal" of the encevis artifact reduction are shown to be non-inferior to the parameters of predicate device Persyst. Moreover in 73 out of 127 test cases, the suppression of clean EEG by encevis was lower compared to Persyst. And in 83 out of 93 test cases, the SNR after artifact removal by encevis was higher compared to Persyst. It can be concluded that the encevis artifact reduction "PureEEG" does not perform worse that the artifact reduction by the predicate device.

encevis rhythmic and periodic patterns performance testing:

The detection of rhythmic and periodic patterns in NeuroTrend is used to visually mark EEG segments with rhythmic or periodic signal content. The definition of rhythmic and periodic patterns follow the guidelines of the ACNS (American Clinical Neurophysiology Society) ICU EEG Terminology (Hirsch et al., 2013). NeuroTrend displays all detected rhythmic and periodic patterns in plots called "Pattern Localization" and "Pattern Frequency".

For the validation we compared and statistically analyze annotations of two human EEG-readers with the detections of NeuroTrend. We showed that the detected patterns have a high sensitivity and specificity compared to manual annotated EEG segments. We prospectively recorded 83 long term EEGs from ICU-patients at two different centers using the international 10-20 electrode system with a sampling rate of 256Hz.

EEGs were annotated by two clinical neurophysiologists that were naive to these EEGs. The annotation procedure included the first minute of each hour, were each minute was split into three independent segments of 20 seconds resulting in 11935 common annotation segments. Several non-overlapping annotations were allowed in each annotation segment. Each annotation may have an arbitrary start and end position but has to be fully included in the annotation segment. For each annotation, the reviewer was allowed to choose between one of the following pattern types:

• 1. PD: periodic pattern
• 2. RDA: rhythmic delta activity
• 3. RTA: rhythmic theta activity
• 4. RAA: rhythmic alpha activity
• 5. RDA+S: rhythmic delta activity with superimposed sharp waves or spikes (RDA+S). (equivalent to SW in Version 1.6)
• 6. BS: burst suppression pattern
• 7. No annotation (short NOPA).

{22}------------------------------------------------

Image /page/22/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in gray, with the "A" stylized as a triangle. To the right of the letters is the text "AUSTRIAN INSTITUTE OF TECHNOLOGY" in a dark red color. The text is stacked on top of each other.

In addition to the type of the pattern the localization property had to be set by the human reviewers. This property was defined in (Hirsch et al., 2013) as main term 1:

• 1. G: generalized pattern
• 2. L: lateralized pattern

The annotations from the two reviewers were then used as gold standard condition to test sensitivity and specificity of the rhythmic and periodic pattern detection of NeuroTrend. Annotations had to be consistent between both reviewers to be used in the sensitivity and specificity measurement.

The detection performance was defined by assigning one of four possible test conditions to each of the 1 minute annotation segments: true positive (TP), true negative (FP), true negative (TN), and false neqative (FN). An annotation segment was counted as TP if a detection and an annotation was present. An annotation segment with a gold standard annotation but without any detection will be counted as FN. An annotation segment with detections but without annotations will be counted as FP. An annotation segment without gold standard annotation and without detections will be counted as TN.

The sensitivity is defined as:

$$\mathbf{SE} \text{ [%]} = #\text{TP} / (#\text{TP} + #\text{FN}) \text{ * } \text{100}$$

The specificity is defined as:

SP [%] = #TN/(#TN+#FP) * 100

The # symbol stands for "number of". The symbol "#TP" represents therefore the number of true positive annotation segments.

The localization information will be validated by comparing the concise annotations of the two human reviewers for all correctly detected markers (the TP detections).

The result of the manual annotation procedure was evaluated using the Cohens' kappa statistic. This statistic measures the level of agreement between two reviewers. A kappa value of 0.66 was measured between reviewer 1 and reviewer 2.

		REVIEWER 2
		NOPAT	PD	RAA	RDA	RTA	RDA+S
REVIEWER 1	NOPAT	10757	311	14	78	23	4
	PD	588	1290	0	63	14	0
	RAA	1	1	6	1	10	0
	RDA	135	5	0	119	1	4
	RTA	50	25	1	23	107	0
	RDA+S	10	0	3	0	0	20
	Cohens Kappa:0.66 (CI=0.64-0.67)			Substantial agreement

Table 3: Cohens' kappa statistic for the evaluation of the pattern detection

The overall detection performance measures the sensitivity of the NeuroTrend detections without evaluating the pattern type. The result is marked with the label "ANY" in the result file. This result proofs the ability of NeuroTrend to detect any relevant pattern and ignores pattern type mismatches. The result of the periodic pattern group is labeled as "PD". This result shows the sensitivity and specificity of the periodic pattern detections. The result of the rhythmic delta activity

{23}------------------------------------------------

Image /page/23/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters AIT in a stylized font, with the letters in gray. To the right of the letters, the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" are written in a smaller, red font. The logo is simple and modern, and it is likely used to represent the organization's brand.

pattern detections is labeled as "RDA". The result of the ARA group shows the result of aggressive rhythmic activity, including the pattern types RTA, RAA, and RDA+S.

Pattern Type	Sensitivity[%]	Specificity[%]
ANY	81.86 (79.9 - 83.8)	83.80 (83.1 - 84.5)
PD	69.73 (67.2 - 72.3)	95.89 (95.5 - 96.3)
ARA (RTA, RAA, RDA+S)	89.40 (84.2 - 94.6)	94.85 (94.5 - 95.3)
RDA	91.73 (86.4 - 97.1)	86.05 (85.4 - 86.7)

Table 4: Sensitivity and specify for encevis pattern detection

The inter reader agreement table of the localization information (ACNS Main Term 1) compares the consistent annotations of two EEG experts to the localization shown in NeuroTrend. The result is shown in the following table:

REVEWER 1+2		NeuroTrend
		G	L
	G	891	86
	L	130	175
Cohens Kappa:		0.51, CI=0.45-0.57(Moderate agreement)

Table 5: Inter-reader agreement between reviewers and NeuroTrend for encevis pattern localization

encevis aEEG performance testing:

Amplitude-integrated EEG (aEEG) is a popular method for monitoring cerebral function by displaying the amplitude trend of brain activity. It is the boundary of the EEG waveform (i.e. the envelope) and not the EEG itself (i.e. the carier) that characterizes the tendency of amplitude changes (Zhang and Ding, 2013).

The Background-AEEG module of NeuroTrend estimates and visualizes the temporal evolution (trend) of the EEG amplitude. The implementation is oriented on the proposed method of (Zhang and Ding, 2013)

In the first step the frequency response of the module is checked for equality with the proposed method of (Zhang and Ding. 2013). This test only considers the correct slope (dB loss per decade) not the correct filter gain factor. In this test, sinusoidal one-channel test data with increasing from 0.5Hz to 32Hz and amplitude of 40uV are generated, one test case for each hemisphere. With the results of the module the frequency response is determined and checked if the dB loss per decade within the band pass (cut-off frequencies of 2 and 15Hz) is -12db/dec and the maximum gain factor in the stop band is not greater than -30dB. This step validates the correct implementation of the filters and its characteristics (expect the gain factor) within the module.

In the second step the results of the module are compared with the aEEG results of Persyst (CE certified and FDA approved software; http://www.persyst.com/) using real EEG data. The configuration of Persyst is set in a way to allow an adequate comparison.

After successful validation according to the description above we will have shown that the Background-AEEG module correctly determines the averaged EEG amplitude of the left and right hemisphere according to the proposes method of (Zhang and Ding, 2013)

510(k) summary – Page 20

{24}------------------------------------------------

Image /page/24/Figure/0 description: The image contains two line graphs showing frequency response. The graph on the left is labeled "Frequency response: C3-P3", while the graph on the right is labeled "Frequency response: C4-P4". Both graphs show a similar trend, with the y-axis representing "20*log(a/a_o) [dB]" and the x-axis representing "f [Hz]", ranging from 1 to 20.

Figure 9: Frequency response of the left (left side) and the right hemisphere (right side).

In the first validation step the frequency response of the aEEG module is checked for equality with the pro-posed method of (Zhang and Ding, 2013). The resulting frequency response of the Background-AEEG module is shown in Figure 9 .The following conclusions were drawn from the results:

• . The determined characteristic is very similar to the published version in (Zhang and Ding, 2013). Only the absolute shift of the complete frequency response is different but because only changes in aEEG values are of clinical relevance this detail is irrelevant.
• Both hemispheres show the same characteristic ●
• . In the stop band there is a suppression of -30dB and higher
• . The slope in the pass band is approximately -12dB/decade

In the second step the results of the aEEG module are compared with the aEEG results of Persyst. For this test, real EEG data were used. The aEEG of the same EEG segment using either Persyst or the Background-AEEG module of encevis NeuroTrend were compared. The test cases showed that the aEEG values of the Background-AEEG module of encevis NeuroTrend and Persyst are in good accordance. Furthermore, the aEEG values are in good accordance with the corresponding raw EEG.

encevis frequency bands performance testing:

The background-frequency module of NeuroTrend estimates and visualizes the temporal evolution (trend) of relative proportions of dominant EEG-waveform-frequencies. The result is graphically presented using a plot (cf. Figure 10), where the x-axis represents the time-axis, and four stacked areas in different colors and widths represent the relative proportions of the four frequency bands Delta, Theta. Alpha, and Beta for subsequent time windows with lengths of 15 seconds. The intensity of the colors furthermore corresponds to the amplitudes in these four frequency bands. This representation allows the user to identify time epochs that are dominated by a specific frequency band. E.g., EEGslowing or, in other words, an epoch with dominant delta- or theta-wave can be recognized in the graphical representation by broad stretches of the corresponding areas.

Image /page/24/Figure/10 description: The image shows a stacked proportion chart of frequency bands. The y-axis ranges from 0% to 100%, and the x-axis represents time. The chart displays the proportions of different frequency bands, including beta, alpha, theta, and delta. The delta band appears to be the most dominant.

Figure 10: Graphical representation of the Background-EEG-Frequency evaluation results.

{25}------------------------------------------------

Image /page/25/Picture/2 description: The image shows the logo for the Austrian Institute of Technology (AIT). The logo consists of the letters "AIT" in a stylized, geometric font, with the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" to the right of the letters. The letters are in gray, while the words are in a dark red color. The logo is simple and modern, and it is likely used to represent the organization's brand.

In order to proof the validity of the Background-EEG-frequency module we followed a twostep approach. In the first step it was shown that the assignment of sinusoidal test data to frequency bands (Delta, Theta, Alpha, or Beta) is correct according to the above definitions of frequency borders. In this test, sinusoidal test data with frequencies across all four bands and amplitudes ranging from 2 uV to 200 µV were generated. Then it was verified, that the algorithm correctly assigns each test signal to the corresponding frequency band, and that the measurement error for amplitudes are below 5 %. This validates the correct assignments of single, 3-second EEG epochs to a frequency band and amplitude.

In a seconds step it is shown that the globally dominant background frequency within a 15-seconds window is correctly identified. This is done using manually selected EEG recordings from epilepsy- or ICU patients. Each of these EEG samples is representative for a specific background-EEG-frequency band, i.e., it is mainly dominated by delta-, theta-, alpha-, or beta-waves. For these samples the background-EEG-frequency module calculates the proportional composition of frequency bands. The one frequency band with the largest proportion can be seen as the globally dominant background frequency, if this proportion is particularly high. Thus it is verified for each of these representative examples that the relative proportion corresponding to the true frequency band is greater than 50 %.

encevis Burst Suppression performance testing:

The detection of burst suppression patterns and quantitative measure for the discontinuity of the EEG shown in NeuroTrend was validated using the following approach:

• 1. The time point of the detected burst suppression patterns will be compared to annotations defined by two clinical EEG experts using EEG data from a multicenter study. Sensitivity and specificity will be calculated.
• 2. The quantitative measure of the amplitude loss of the suppression and the suppression time in percent will be validated using an artificial EEG. The EEG file includes a set burst suppression patterns with different values for suppression time and suppression amplitude loss. The calculation results of the quantitative burst suppression plots shown in NeuroTrend will be compared to precalculated values.

We recorded 83 long term EEGs from intensive care patients from two different centers using the international 10-20 electrode system with a sampling rate of 256Hz. EEGs were annotated by two clinical neurophysiologists that were naive to these EEGs. The annotation procedure included the first minute of each hour resulting in 3978 valid annotation segments. The reviewers were allowed to assign two categories for each annotation segment:

• 8. EEG with burst suppression patterns (BS)
• 9. EEG without burst suppression patterns (BS)

Statistical analysis of the detection performance was done by defining the annotations of the reviewers as gold standard and by comparing these annotations to the detection results of the computational method. Each one minute EEG segments annotated as "EEG with burst suppression" with an overlapping burst suppression detection segment of 15 seconds was defined as true positive (TP) event. One minute EEG segments annotated as "EEG with burst suppression patterns" without any overlapping burst suppression detection result were defined as false negatives (FN). One minute seqments annotated as "EEG without burst suppression patterns" and with an overlapping burst suppression detection result are defined as false positives (FP), all other segments are defined as true negatives (TN).

{26}------------------------------------------------

Image /page/26/Picture/2 description: The image shows the logo for the Austrian Institute of Technology. The logo consists of the letters "AIT" in a stylized font, with the words "AUSTRIAN INSTITUTE OF TECHNOLOGY" written in smaller letters to the right of the letters. The letters and words are in a dark gray color. The background is white.

The following table shows the evaluation results of the automatic burst suppression detection method in NeuroTrend using 3978 segments annotated by two reviewers. The results of the automatic burst suppression detection method were compared to the manual annotations of the reviewers. The detection performance was analyzed for consensus annotations of the two reviewers. The consensus annotations only include annotation segments where both reviewers showed the same decision about Burst Suppression pattern. The measured values for sensitivity (SE), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) prove the validly of the detection algorithm. The very large sample size does imply a statistically high confidence.

Rev. (n)	SE (%)	SP (%)	PPV (%)	NPV (%)
2	87(84.7-89.9)	92(91.4-92.9)	61(57.9-64.3)	98(97.7-98.5)

Table 6: Performance of the automatic burst suppression detection method

10. Statement of Substantial Equivalence

encevis is substantially equivalent in design and intended use to the predicate device. Any differences between the subject and predicate device have no significant influence on safety or effectiveness as established through performance testing. Therefore, the encevis raises no new issues of safety or effectiveness when compared to the predicate device.

Regulation Number and Section

§ 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).