(161 days)
EnsoSleep is intended for use in the diagnostic evaluation by a physician to assess sleep quality and as an aid for physicians in the diagnosis of sleep disorders and respiratory related sleep disorders in pediatric as follows:
- · Pediatric patients 13 years and older with polysomnography (PSG) tests obtained in a Hospital or Sleep Clinic
- · Adult patients with PSGs obtained in a Hospital or Sleep Clinic
- · Adult patients with Home Sleep Tests
EnsoSleep is a software-only medical device to be used under the supervision of a clinician to analyze physiological signals and automatically score sleep study results, including the staging of arousals, leg movements, and sleep disordered breathing events including obstructive apneas (OSA), central sleep apneas (CSA), and hypopneas.
All automatically scored events and physiological signals which are retrieved, analyzed, displayed, and summarized are subject to verification by a qualified clinician. Central sleep apneas (CSA) should be manually reviewed and modified as appropriate by a clinician.
All events can be manually marked or edited within records during review.
Photoplethysmography (PPG) total sleep time is not intended for use when electroencephalograph (EEG) data is recorded. PPG total sleep time is not intended to be used as the sole or primary basis for diagnosing any sleep related breathing disorder, prescribing treatment, or determining whether additional diagnostic assessment is warranted.
EnsoSleep is a software-only medical device that analyzes previously recorded physiological signals obtained during sleep. Users of EnsoSleep are consistent with the roles required to run a sleep clinic: sleep physicians, sleep technicians, clinic operations managers, and IT administrators. EnsoSleep can analyze at-home and in-lab sleep studies for both adult and pediatric patients who are at least 13 years old. Automated algorithms are applied to the raw signals in order to derive additional signals and interpret the raw and derived signal information. The software automates recognition of the following: respiratory events, sleep staging events, arousal events, movement events, cardiac events, derived signals, and calculated indices. EnsoSleep does not interpret the results, nor does it suggest a diagnosis. The device only marks events of interest for review by a physician who is responsible for diagnoses. The device does not analyze data that are different from those analyzed by human scorers.
The signals and automated analyses can be visually inspected and edited prior to the results being integrated into a sleep study report.
The software consists of 4 major components:
- The Application Platform runs on local clinic workstations and manages the detection, upload, and download of study records and scoring to and from the Storage Platform
- The Processing Platform accepts raw physiological signals as inputs in order to recognize events, derive signals, and calculate indices
- . The Storage Platform facilitates file and database storage in the EnsoSleep cloud through an API
- The Dashboard is a web-based user interface to support configuration, clinic management, and sleep study scoring
Here's a breakdown of the acceptance criteria and the study that proves the device meets them, based on the provided text:
1. Table of Acceptance Criteria and Reported Device Performance
The text describes that "the acceptance criteria were selected based on PA, NA, OA, MAE, Deming Regression coefficient, Bland-Altman mean difference and limits of agreement performance criteria that validate performance substantially equivalent to, or greater than, but not lesser than by more than 10% or similarly defined criteria in any category of the predicate 510(k) reported device performance across all endpoints respectively."
For detailed performance metrics, the document presents tables for specific endpoints. Below is a structured representation focusing on the key performance indicators mentioned and a comparison to the predicate device where available.
Endpoint 1: Sleep Staging Event Detection (EnsoSleep K210034 vs. Predicate K162627)
| Metric | Acceptance Criteria (Conceptual) | EnsoSleep (K210034) Adult Sample (Reported) | EnsoSleep (K210034) Pediatric Sample (Reported) | Predicate (K162627) Adult Sample (Reported) |
|---|---|---|---|---|
| Positive Agreement (PA) | ≥ Predicate PA - 10% (Ideally ≥ Predicate PA) | Wake: 93.5% | Wake: 93.1% | Wake: 86% |
| N1: 37.0% | N1: 43.2% | N1: 41% | ||
| N2: 88.3% | N2: 92.6% | N2: 77% | ||
| N3: 80.0% | N3: 92.3% | N3: 81% | ||
| REM: 90.9% | REM: 80.9% | REM: 79% | ||
| Negative Agreement (NA) | ≥ Predicate NA - 10% (Ideally ≥ Predicate NA) | Wake: 97.2% | Wake: 99.2% | Wake: 97% |
| N1: 98.3% | N1: 98.8% | N1: 94% | ||
| N2: 89.3% | N2: 89.4% | N2: 87% | ||
| N3: 96.3% | N3: 97.5% | N3: 93% | ||
| REM: 99.3% | REM: 99.1% | REM: 99% | ||
| Overall Agreement (OA) | ≥ Predicate OA - 10% (Ideally ≥ Predicate OA) | Wake: 96.1% | Wake: 97.9% | Wake: 94% |
| N1: 95.0% | N1: 97.0% | N1: 91% | ||
| N2: 88.8% | N2: 90.9% | N2: 83% | ||
| N3: 95.0% | N3: 96.6% | N3: 92% | ||
| REM: 98.3% | REM: 97.0% | REM: 96% | ||
| Total (Overall) OA (Pooled) | ≥ Predicate Total OA - 10% (Ideally ≥ Predicate Total OA) | Adult: 86.6% | Pediatric: 89.7% | Adult: 78% |
| Conclusion on Acceptance Criteria (Endpoint 1) | All 3 EnsoSleep PA, NA, and OA point-estimates vs 2/3 Majority Scoring were observed to be greater than the predicate device PA, NA, and OA point-estimates in some events, with statistically significant results in terms of higher agreement for several stages (e.g., Adult REM, Pediatric Wake, N3, Total). None were observed 10% or lower than 2/3 Majority. |
Endpoint 2: Sleep Apnea Diagnostic Agreement (Per-Patient AHI)
| Metric | Acceptance Criteria (Conceptual) | EnsoSleep (K210034) Adult Sample (Reported) | EnsoSleep (K210034) Pediatric Sample (Reported) | Predicate (K162627) Adult Sample (Reported) |
|---|---|---|---|---|
| Positive Percent Agreement (PA) | ≥ Predicate PA - 10% (Ideally ≥ Predicate PA) | AHI ≥ 5: 94.4% | AHI ≥ 1: 94.4% | AHI ≥ 5: 91% |
| AHI ≥ 15: 94.0% | AHI ≥ 5: 90.5% | AHI ≥ 15: 95% | ||
| REM AHI ≥ 5: 86.7% | AHI ≥ 10: 78.6% | REM AHI ≥ 5: 83% | ||
| REM AHI ≥ 15: 81.5% | AHI ≥ 15: 85.7% | REM AHI ≥ 15: 79% | ||
| Negative Percent Agreement (NA) | ≥ Predicate NA - 10% (Ideally ≥ Predicate NA) | AHI ≥ 5: 89.7% | AHI ≥ 1: 77.8% | AHI ≥ 5: 76% |
| AHI ≥ 15: 96.3% | AHI ≥ 5: 100.0% | AHI ≥ 15: 98% | ||
| REM AHI ≥ 5: 83.0% | AHI ≥ 10: 94.9% | REM AHI ≥ 5: 89% | ||
| REM AHI ≥ 15: 93.3% | AHI ≥ 15: 100.0% | REM AHI ≥ 15: 96% | ||
| Overall Percent Agreement (OA) | ≥ Predicate OA - 10% (Ideally ≥ Predicate OA) | AHI ≥ 5: 93.0% | AHI ≥ 1: 89.4% | AHI ≥ 5: 85% |
| AHI ≥ 15: 95.0% | AHI ≥ 5: 95.7% | AHI ≥ 15: 97% | ||
| REM AHI ≥ 5: 85.0% | AHI ≥ 10: 91.5% | REM AHI ≥ 5: 86% | ||
| REM AHI ≥ 15: 90.0% | AHI ≥ 15: 97.9% | REM AHI ≥ 15: 92% | ||
| Conclusion on Acceptance Criteria (Endpoint 2) | EnsoSleep PA, NA, and OA point-estimates vs 2/3 Majority Scoring were observed to be greater than the predicate device for some OSA severity categories in both adult and pediatric samples. Only one instance (Pediatric AHI > 10 PA) was within 10% of the predicate. All met or exceeded objective performance goals. |
Endpoint 3: Sleep Scoring Event Detection (EnsoSleep K210034 vs. Predicate K162627 and Reference K112102)
| Metric | Acceptance Criteria (Conceptual) | EnsoSleep (K210034) Adult Sample (Reported) | EnsoSleep (K210034) Pediatric Sample (Reported) | Predicate (K162627) Adult Sample (Reported) | Reference (K112102) Adult Sample (Reported) |
|---|---|---|---|---|---|
| Positive Agreement (PA) | ≥ Reference PA - 10% (Ideally ≥ Reference PA) | SDB: 75.4% | SDB: 72.7% | SDB: 67% | N/A (for SDB) |
| Hypopnea: 66.3% | Hypopnea: 68.8% | Hypopnea: 60.3% | Hypopnea: 60.3% | ||
| Obstructive Apnea: 74.1% | Obstructive Apnea: 45.5% | Obstructive Apnea: 53% | N/A (for Obstructive Apnea) | ||
| Central Apnea: 65.3% | Central Apnea: 68.9% | Central Apnea: 63.8% | Central Apnea: 63.8% | ||
| Arousal: 73.6% | Arousal: 78.6% | Arousal: 66% | N/A (for Arousal) | ||
| Leg Movement: 82.0% | Leg Movement: 66.0% | Leg Movement: 71% | N/A (for Leg Movement) | ||
| Negative Agreement (NA) | ≥ Reference NA - 10% (Ideally ≥ Reference NA) | SDB: 97.0% | SDB: 98.6% | SDB: 93% | N/A (for SDB) |
| Hypopnea: 97.1% | Hypopnea: 98.9% | Hypopnea: 97.6% | Hypopnea: 97.6% | ||
| Obstructive Apnea: 99.3% | Obstructive Apnea: 99.7% | Obstructive Apnea: 97% | N/A (for Obstructive Apnea) | ||
| Central Apnea: 99.5% | Central Apnea: 99.7% | Central Apnea: 99.6% | Central Apnea: 99.6% | ||
| Arousal: 95.6% | Arousal: 97.0% | Arousal: 90% | N/A (for Arousal) | ||
| Leg Movement: 92.4% | Leg Movement: 95.5% | Leg Movement: 90% | N/A (for Leg Movement) | ||
| Overall Agreement (OA) | ≥ Reference OA - 10% (Ideally ≥ Reference OA) | SDB: 94.9% | SDB: 97.6% | SDB: 91% | N/A (for SDB) |
| Hypopnea: 95.5% | Hypopnea: 98.0% | Hypopnea: N/R | Hypopnea: N/R | ||
| Obstructive Apnea: 98.8% | Obstructive Apnea: 99.5% | Obstructive Apnea: 96% | N/A (for Obstructive Apnea) | ||
| Central Apnea: 98.9% | Central Apnea: 99.5% | Central Apnea: N/R | Central Apnea: N/R | ||
| Arousal: 93.2% | Arousal: 95.5% | Arousal: 87% | N/A (for Arousal) | ||
| Leg Movement: 91.7% | Leg Movement: 94.5% | Leg Movement: 89% | N/A (for Leg Movement) | ||
| Conclusion on Acceptance Criteria (Endpoint 3) | EnsoSleep PA, NA, and OA point-estimates observed to be greater than or within 5% of the reference device performance for all event types, with statistically significant differences (greater performance) in a majority of cases. All met or exceeded objective performance goals. |
Endpoint 4: Total Sleep Time (TST) and Respiratory Rate (RR)
| Metric | Acceptance Criteria (Conceptual) | EnsoSleep PPG-TST (K210034) RR Sample (Reported) | EnsoSleep EEG-TST (K210034) RR Sample (Reported) | EnsoSleep EEG-TST (K210034) Adult Sample (Reported) | EnsoSleep EEG-TST (K210034) Pediatric Sample (Reported) |
|---|---|---|---|---|---|
| Deming Regression Slope ($β$1) | Near unity (0.90 < $β$1 < 1.10) | 0.964 | 0.984 | 1.037 | 1.006 |
| Deming Regression Intercept ($β$0) [hours] | Near zero ($β$0 < 15 minutes / e.g., < 0.25 hours) | 0.089 | 0.156 | -0.181 | 0.021 |
| Bland-Altman Mean Difference (MD) [minutes] | Near zero (MD within ≤15 minutes) | 5.380 | -4.785 | 0.515 | -3.255 |
| Bland-Altman Upper Limit of Agreement (ULOA) [min] | Within <90 minutes | 73.463 | 32.922 | 57.750 | 10.654 |
| Bland-Altman Lower Limit of Agreement (LLOA) [min] | Within <90 minutes | -62.703 | -42.492 | -56.720 | -17.164 |
| RR Performance | ≥ 90% percent epochs within ≤2 brpm of reference; MAE ≤ 2 brpm of reference | Met acceptance criteria (no specific numerical values provided in the table) | Met acceptance criteria (no specific numerical values provided in the table) | N/A | N/A |
| Conclusion on Acceptance Criteria (Endpoint 4) | EnsoSleep PPG-TST and EEG-TST demonstrated statistically similar performance to the predicate device in all samples, meeting or exceeding all acceptance criteria for Deming regression and Bland-Altman analysis, demonstrating no clinically significant deviations. |
Overall Conclusion on Study Meeting Acceptance Criteria:
"The subject EnsoSleep device event detection and diagnostic agreement performance were observed to meet or exceed the PA, NA, and OA performance acceptance in the 26 total experiments (26/26) across all 4 experimental endpoints evaluated..." indicating that the device successfully met all predefined acceptance criteria.
2. Sample Sizes Used for the Test Set and Data Provenance
- Test Set Sample Sizes:
- Adult Sample: N=100 adult subjects (for Sleep Staging, Sleep Apnea Diagnostic, Event Detection, and TST/RR)
- Pediatric Sample: N=47 pediatric subjects (for Sleep Staging, Sleep Apnea Diagnostic, and Event Detection, and TST/RR)
- Respiratory Rate (RR) Sample: N=100 adult subjects (specifically for Respiratory Rate and TST analysis)
- Data Provenance:
- "Archived collection of retrospective diagnostic clinical PSG subject data."
- The data was collected from "five (5) clinical testing laboratories" where "two (2) AASM Accredited Sleep Testing Facilities were selected each with two (2) regional sleep testing centers."
- No specific country of origin is mentioned, but the regulatory submission is to the FDA (U.S. Food & Drug Administration). Given the use of AASM (American Academy of Sleep Medicine) accreditation, it is highly likely the data originated from the United States.
- The data is explicitly stated as retrospective.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of Those Experts
- Number of Experts: A panel of 9 total registered scoring technologists (RPSGTs) was used to establish the ground truth. From this panel, any given subject was assigned to 3 additional, prospective scorers.
- Qualifications of Experts: The RPSGTs had "5 to 20+ years clinical experience". They were verified to meet "all defined study scoring-acquisition, scoring-blind, and rater-quality certification controls."
4. Adjudication Method for the Test Set
- The ground truth was established using a "2/3 Majority Scoring consensus reference". This means that for any given event or stage, at least two out of the three independent scorers had to agree for it to be considered part of the ground truth.
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
- The study described is not a Multi-Reader Multi-Case (MRMC) comparative effectiveness study evaluating human readers' improvement with AI assistance.
- Instead, it's a standalone performance study comparing the EnsoSleep algorithm's performance against expert consensus (human-scored ground truth) and demonstrating substantial equivalence to a predicate device.
- The study design focused on the agreement of the device with human scorers, not on how the device assists human readers or changes their performance.
6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) was Done
- Yes, a standalone performance study was done. The described tests for "event detection and diagnostic agreement performance" of the EnsoSleep device relate to its direct output compared to the ground truth established by human experts. The system "automatically score sleep study results" and "does not interpret the results, nor does it suggest a diagnosis." It marks events for a clinician's review and verification, implying its standalone detection capability is what was assessed.
7. The Type of Ground Truth Used (Expert Consensus, Pathology, Outcomes Data, etc.)
- The type of ground truth used was expert consensus, specifically a "2/3 Majority Scoring consensus reference" established by independent, registered sleep technologists (RPSGTs).
8. The Sample Size for the Training Set
- The document does not explicitly state the sample size for the training set. The sample sizes provided (N=100 Adult, N=47 Pediatric, N=100 RR) refer specifically to the test sets used for clinical validation, described as "study sample[s]... to construct the final study sample."
9. How the Ground Truth for the Training Set Was Established
- Since the training set size and characteristics are not detailed, the method for establishing its ground truth is also not described in this document. It is implied that the algorithm was trained on prior data, but the specifics of that process are outside the scope of this 510(k) summary, which focuses on the validation of the final product.
{0}------------------------------------------------
Image /page/0/Picture/0 description: The image contains the logo of the U.S. Food and Drug Administration (FDA). The logo consists of two parts: the Department of Health & Human Services logo on the left and the FDA text logo on the right. The FDA text logo is in blue and includes the letters "FDA" followed by the words "U.S. Food & Drug Administration".
EnsoData, Inc. Sigrid Schoepel Director of Regulatory Affairs 111 S. Hamilton St. Suite 30 Madison, Wisconsin 53703
Re: K210034
Trade/Device Name: EnsoSleep Regulation Number: 21 CFR 882.1400 Regulation Name: Electroencephalograph Regulatory Class: Class II Product Code: OLZ Dated: May 11, 2021 Received: May 12, 2021
Dear Sigrid Schoepel:
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 801); medical device reporting of medical device-related adverse events) (21 CFR 803) for
{1}------------------------------------------------
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 medical devices and radiation-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 (DICE@fda.hhs.gov) or phone (1-800-638-2041 or 301-796-7100).
Sincerely,
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) K210034
Device Name EnsoSleep
Indications for Use (Describe)
EnsoSleep is intended for use in the diagnostic evaluation by a physician to assess sleep quality and as an aid for physicians in the diagnosis of sleep disorders and respiratory related sleep disorders in pediatric as follows:
- · Pediatric patients 13 years and older with polysomnography (PSG) tests obtained in a Hospital or Sleep Clinic
- · Adult patients with PSGs obtained in a Hospital or Sleep Clinic
- · Adult patients with Home Sleep Tests
EnsoSleep is a software-only medical device to be used under the supervision of a clinician to analyze physiological signals and automatically score sleep study results, including the staging of arousals, leg movements, and sleep disordered breathing events including obstructive apneas (OSA), central sleep apneas (CSA), and hypopneas.
All automatically scored events and physiological signals which are retrieved, analyzed, displayed, and summarized are subject to verification by a qualified clinician. Central sleep apneas (CSA) should be manually reviewed and modified as appropriate by a clinician.
All events can be manually marked or edited within records during review.
Photoplethysmography (PPG) total sleep time is not intended for use when electroencephalograph (EEG) data is recorded. PPG total sleep time is not intended to be used as the sole or primary basis for diagnosing any sleep related breathing disorder, prescribing treatment, or determining whether additional diagnostic assessment is warranted.
Type of Use (Select one or both, as applicable)
| ☑ Prescription Use (Part 21 CFR 801 Subpart D) |
|---|
| ☐ Over-The-Counter Use (21 CFR 801 Subpart C) |
CONTINUE ON A SEPARATE PAGE IF NEEDED.
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 PRAStaff(@fda.hhs.gov
"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."
{3}------------------------------------------------
K210034 EnsoSleep Traditional 510(k) Summary Prepared in accordance with the content and format outlined in 21 CFR 807.92
T SUBMITTER INFORMATION
| Name: | EnsoData, Inc. |
|---|---|
| Address: | 111 S. Hamilton Street, Suite 30 |
| Madison, WI 53703 USA | |
| Phone: | (608)509-4704 |
| Contact Person: | Sigrid Schoepel |
| Email: | sigrid@ensodata.com |
| Date: | June 8, 2021 |
II. SUBJECT DEVICE INFORMATION
| Trade Name: | EnsoSleep |
|---|---|
| Common Name: | Automatic Event Detection Software for Polysomnography with Electroencephalograph (EEG) |
| Classification Name: | Electroencephalograph 21 CFR 882.1400 |
| Regulatory Class: | II |
| Product Code: | OLZ |
| Intended Use: | EnsoSleep is intended for use for the diagnostic evaluation by a physician to assess slee |
leep quality and as an aid for the diagnosis of sleep disorders and respiratory related sleep disorders in pediatric patients ages 13 years and older. EnsoSleep is a software-only medical device to be used under the supervision of a clinician to analyze physiological signals and automatically score sleep study results, including the staging of sleep, detection of arousals, leg movements, and sleep disordered breathing events including obstructive apneas, central apneas, and hypopneas. All automatically scored events and physiological signals analyzed are retrieved, displayed, and summarized, and are subject to verification by a qualified clinician. All events can be manually marked or edited within records during review.
Target Patient Population: Adult and Pediatric
III. PREDICATE DEVICE K162627 Trade Name: EnsoSleep Target Patient Population: Adults only
{4}------------------------------------------------
SUBJECT DEVICE DESCRIPTION IV.
EnsoSleep is a software-only medical device that analyzes previously recorded physiological signals obtained during sleep. Users of EnsoSleep are consistent with the roles required to run a sleep clinic: sleep physicians, sleep technicians, clinic operations managers, and IT administrators. EnsoSleep can analyze at-home and in-lab sleep studies for both adult and pediatric patients who are at least 13 years old. Automated algorithms are applied to the raw signals in order to derive additional signals and interpret the raw and derived signal information. The software automates recognition of the following: respiratory events, sleep staging events, arousal events, movement events, cardiac events, derived signals, and calculated indices. EnsoSleep does not interpret the results, nor does it suggest a diagnosis. The device only marks events of interest for review by a physician who is responsible for diagnoses. The device does not analyze data that are different from those analyzed by human scorers.
The signals and automated analyses can be visually inspected and edited prior to the results being integrated into a sleep study report.
The software consists of 4 major components:
- The Application Platform runs on local clinic workstations and manages the detection, ● upload, and download of study records and scoring to and from the Storage Platform
- The Processing Platform accepts raw physiological signals as inputs in order to recognize events, derive signals, and calculate indices
- . The Storage Platform facilitates file and database storage in the EnsoSleep cloud through an API
- The Dashboard is a web-based user interface to support configuration, clinic management, and sleep study scoring
Input signals used to derive outputs include:
- Electroencephalogram (EEG) ●
- Electrocardiogram (ECG) ●
- Electro-oculogram (EOG) ●
- Electromyogram (EMG) ●
- Actigraphy
- Airflow
- Oximetry
- Saturation of Peripheral Oxygen (SpO2)
- Respiratory Inductance Plethysmogram (RIP)
- Polyvinylidene Flouride (PVDF)
- Photoplethysmogram (PPG)
- Pulse Rate
- Snoring Microphone ●
- . Esophogeal Manometry
Outputs that are displaved to users include:
- Respiratory Events
- o Obstructive Sleep Apneas (OSA)
{5}------------------------------------------------
- Central Sleep Apneas (CSA) o
- Mixed Sleep Apneas (MSA) O
- Hypopneas O
- Cheyne-Stokes Respiration O
- Periodic Breathing O
- Sleep Staging Events ●
- Stage Wake O
- Stage N1 O
- Stage N2 O
- Stage N3 O
- O Stage REM (rapid eye movement)
- Arousal Events ●
- O Arousals
- Movement Events ●
- o Leg Movements (LM)
- o Periodic Leg Movement Series (PLMS)
- Cardiac Events* ●
- o Bradycardia
- o Tachycardia
- Respiratory Rate Events
- o Respiratory Rate
- Sleep-Wake Events ●
- o Wake
- o Sleep
- Apnea-Hypopnea Index (AHI) ●
- Sleep Architecture ●
- Sleep Efficiency (SE) ●
- Arousal Index (ArI) ●
- Sleep Latency (SL) .
- REM Latency (RL)
- Total Sleep Time (TST)
- Periodic Leg Movements (PLMS) Index ●
Capnogram data are not displayed to users.
The primary differences between the predicate device and subject device include:
-Target patient population: pediatric patients are included
-Automated scoring of Cheyne-Stokes and Periodic Breathing respiration events and Total Sleep Time
-Browser-based scoring interface to view and edit sleep studies
-Home-sleep studies are supported
- The tachycardia and bradycardia outputs are not for use for cardiovascular monitoring or diagnosis, nor does the device detect arrhythmias.
{6}------------------------------------------------
V. COMPARISON OF TECHNOLOGICAL CHARACTERISTICS WITH THE PREDICATE DEVICE
The table below compares the predicate device EnsoSleep with the subject device EnsoSleep.
| Element | Predicate DeviceEnsoData, Inc.K162627EnsoSleep | Submitted DeviceEnsoData, Inc.K210034EnsoSleep | Discussion |
|---|---|---|---|
| Indications for use | EnsoSleep is intended for usefor the diagnostic evaluationby a physician to assess sleepquality and as an aid for thediagnosis of sleep andrespiratory related sleepdisorders in adults only.EnsoSleep is a software-onlymedical device to be usedunder the supervision of aclinician to analyzephysiological signals andautomatically score sleepstudy results, including thestaging of sleep, detection ofarousals, leg movements, andsleep disordered breathingevents including obstructiveapneas. All automaticallyscored events are subject toverification by a qualifiedclinician. Central apneas,mixed apneas, and hypopneasmust be manually markedwithin records. | EnsoSleep is intended foruse in the diagnosticevaluation by a physicianto assess sleep quality andas an aid for physicians inthe diagnosis of sleepdisorders and respiratoryrelated sleep disorders inpediatric and adult patientsas follows:• Pediatric patients ages 13years and older withpolysomnography (PSG)tests obtained in a Hospitalor Sleep Clinic• Adult patients with PSGsobtained in a Hospital orSleep Clinic• Adult patients withHome Sleep TestsEnsoSleep is a software-only medical device to beused under the supervisionof a clinician to analyzephysiological signals andautomatically score sleepstudy results, including thestaging of sleep, detectionof arousals, legmovements, and sleepdisordered breathingevents includingobstructive apneas (OSA),central sleep apneas(CSA), and hypopneas.Central sleep apneas(CSA) should be manuallyreviewed and modified asappropriate by a clinician. | Different The subjectdevice includes expansionof the indications for use tothe pediatric population.Clinical testing using thescoring methodology forthe predicate device wereapplied and validation todemonstrate scoring forexpanded uses to pediatricpopulation and newautomated outputs. |
| Element | Predicate DeviceEnsoData, Inc.K162627EnsoSleep | Submitted DeviceEnsoData, Inc.K210034EnsoSleep | Discussion |
| All events can bemanually marked or editedwithin records duringreview. | |||
| Photoplethysmography(PPG) total sleep time isnot intended for use whenelectroencephalograph(EEG) data is recorded.PPG total sleep time is notintended to be used as thesole or primary basis fordiagnosing any sleeprelated breathing disorder,prescribing treatment, ordetermining whetheradditional diagnosticassessment is warranted. | |||
| Intended Use | Analyze pre-recordedphysiological data acquiredduring sleep. | Same as predicate. | Same |
| Patientpopulation | Adults only. | Adults and pediatricpatients 13 years old andabove. | Different |
| Environment ofuse | Physician office (data analysisand reporting). No limitationon where data are acquired. | Same as predicate. | Different; the subjectdevice includes analysis ofhome sleep testing (HST)of adult patients. |
| Signals analyzed | EEG, ECG, EOG, EMGwaveforms; SpO2; respiratoryeffort; airflow; heart/pulserate; snoring loudness; headmovement and position. | EEG, ECG, EOG, EMGwaveforms; SpO2;respiratory effort; airflow;heart/pulse rate; snoringloudness; head movementand position. | Similar |
| Sleep measures | Sleep, REM and N3 onset;total sleep and recordingtimes; sleep efficiency % timeby sleep stage; awakeningsper hour; wake after sleeponset. | Same as predicate. | Same |
| Automaticallyscore sleep stages | Yes; automatically detectsstage Wake (W), REM (R), | Same as predicate. | Same |
| Element | Predicate DeviceEnsoData, Inc.K162627EnsoSleep | Submitted DeviceEnsoData, Inc.K210034EnsoSleep | Discussion |
| NREM 1 (N1), NREM 2 (N2)and slow wave sleep (N3). | |||
| Automaticallyscore sleepdisorderedbreathing events | No; OSA, CSA and mixedhypopneas must be manuallyscored | Yes; automatic detectionof the following events:Obstructive apneas(OSA), Centralapneas (CSA), Hypopneas, Cheyne-Stokesrespiration (CSR), and Periodic breathing(PBE). Additional apnea andsleep disorderedbreathing event typesmay be manuallymarked within therecords. CSA eventsshould also bemanually scored | SimilarSleep disordered breathingevents are detected, and theadditional events areidentified using the existingairflow inputs as in thepredicate device. |
| Automaticallyscore arousalevents | Yes; automatically detectsarousal events. | Yes; automatically detectsarousal events;respiratory-effort relatedarousals, limb movementrelated arousals, andspontaneous corticalarousal. | DifferentThe general arousal eventsdetected by the subjectdevice are now identifiedmore specifically as to thetype of arousal events(same events as detected bythe predicate). |
| Automaticallyscore movementevents | Yes; automatically detects legmovement events. | Same as predicate. | Same |
| Automaticallyscore cardiacevents | No. | Yes; brachycardia andtachycardia. | DifferentThe ECG signals used todetermine these eventswere used in the predicatedevice to determinebreathing, arousal, andmovement events are alsoused to detect cardiacevents. The tachycardia andbradycardia feature is notintended for use forcardiovascular monitoringor diagnosis, nor does the |
| Element | Predicate DeviceEnsoData, Inc.K162627EnsoSleep | Submitted DeviceEnsoData, Inc.K210034EnsoSleep | Discussion |
| Automaticallydeterminederived values | No. | Yes; effort belt respiratoryrate and PPG respiratoryrate. | subject device detectarrhythmias.DifferentThe RIP and PVDF signalsused to determine thesevalues are used in thepredicate device todetermine breathing,arousal, and movementevents. |
| Automated studyuploadanddownload | Yes. | Same as predicate. | Same |
| Automaticallyinitiates studyscoring | Yes. | Same as predicate. | Same |
| Heart rateaccuracy | No. | Same as predicate. | Same |
| Headposition | No. | Same as predicate. | Same |
| Snoring level | No. | Same as predicate. | Same |
| Allows editing insleep studyviewers | Yes. Sleep studies can beopened and edited in sleepstudy scoring devices thatsupport EDF formatted files. | Yes, and within EnsoSleepusing EDF formatted files. | DifferentThe subject device allowsthat signals be edited withinthe existing patient view ofthe web dashboard. |
| Sleep studyreporting | Yes. | Same as predicate. | Same |
| Two-nightreports | No. | Same as predicate. | Same |
| Diseasemanagementcomments | No. | Same as predicate. | Same |
| Data format | EDF | Same as predicate. | Same |
| Compatibility | Operates on any PC withWindows 7 and 8 operatingsystem platforms. | Operates on any PC withWindows 7+, WindowsServer 2012+ operatingsystem platforms. | Updated; Tested on newerversions of Windows asthey are released, using the |
| Element | Predicate DeviceEnsoData, Inc.K162627EnsoSleep | Submitted DeviceEnsoData, Inc.K210034EnsoSleep | Discussion |
| same verificationmethodology. | |||
| Cybersecurity | Authentication controls,authorization controls,cryptographic controls, accesscontrols, checksum controls,software distribution controls,intrusion detection systemcontrols, network and systemscontrols, and databasecontrols. | Same as predicate. | Same |
| Networkrequirements | High-speed internetconnection, above 200 kb/srecommended. | Same as predicate. | Same |
{7}------------------------------------------------
{8}------------------------------------------------
{9}------------------------------------------------
{10}------------------------------------------------
{11}------------------------------------------------
VI. PERFORMANCE DATA
The following performance data were provided in support of the substantial equivalence determination. Since this is a software-only medical device that does not control other devices the performance data do not include biocompatibility, electrical safety, electromagnetic compatibility, mechanical, acoustic, or animal testing.
The intended use population, study population, conditions of interest, designated comparative reference, designated comparative benchmarks, and experimental endpoints were defined. Following the application of clinical laboratory selection controls, five (5) clinical testing laboratories were evaluated, and two (2) AASM Accredited Sleep Testing Facilities were selected each with two (2) regional sleep testing centers that met all laboratory quality, external validity, and subject spectrum controls. Following a cross-sectional study design, an archived collection of retrospective diagnostic clinical PSG subject data was collected and verified to meet the specified disease spectrum, medical condition, medication, and demographic requirements. To construct the final study sample from the archived collection population, a randomized sampling with proportionate allocation across each sleep apnea disease severity quantile (normative, mild, moderate, and severe sleep apnea) was used to construct a valid Adult Sample of N=100 adult subjects (AHI mean: 22.4 [95% Cl: 18.6%, 26.1%], std: 18.8, median: 17.2, min: 0, max: 109.2, rule: 1.a.) and Pediatric Sample of N=47 pediatric subjects (AHI mean: 11.4 [95% CI: 7.6%, 15.6%], std: 13.6, median: 7.0, min: 0, max: 60.3, rule: 1.a.) from the first laboratory N=1984 archived collection of retrospective diagnostic clinical PSG data, and a valid Respiratory Rate Sample of N=100 adult subjects (AHI mean: 13.2 [95% Cl: 9.7%, 16.5%], std: 17.1, median: 6.6, min: 0, max: 82.9, rule: 1.b.) from the second laboratory N=1079 archived collection of retrospective diagnostic clinical PSG data.
The N=100 Adult, N=47 Pediatric, and N=100 Respiratory Rate Study Samples were observed and verified to have no statistically significant differences in the sleep apnea disease state distributional characteristics relative to the laboratory archived collection populations, confirming preservation of disease spectrum breadth and representativeness with respect to the intended use population. The N=100 study samples were verified to contain subjects from all (4/4) sleep apnea disease state severity quantiles, including normative, mild, moderate, and severe sleep apnea. subjects from all (7/7) of the predetermined relevant and/or confounding medical condition groups of interest, including sleep disorders, psychiatric disorders, neurologic disorders, neuro-developmental disorders, cardiac disorders, pulmonary disorders, metabolic and other disorders, subjects from all (5/5) predetermined relevant and/or confounding medication groups of interest, including benzodiazepines, antidepressants, stimulants, opiates, and sleep aids, subjects from each of (7/7) identified demographic group of interest, including all sexes, adult age groups, pediatric age groups, BMI groups, weight groups, and sleepiness groups. Based on these verification and control procedures, the study sample was determined to be a representative sample of the defined intended use and user population. Following the finalization of each N=100 subject Adult, N=47 Pediatric, and N=100 Respiratory Rate study sample respectively, six (6) clinical testing laboratories were evaluated for independent manual scoring, three (3) laboratories were selected, and the clinical test setting was established for constructing a valid 2/3 Majority Scoring consensus reference, using N=9 total registered scoring technologists (RPSGTs) with 5 to 20+ years clinical experience by verification
{12}------------------------------------------------
of meeting all defined study scoring-acquisition, scoring-blind, and rater-quality certification controls. Manual scoring for sleep stage, hypopnea events, obstructive sleep apnea (OSA) events, central sleep apnea (CSA) events, arousal events, limb movement events, respiratory effort related arousal (RERA) events, cheyne-stokes (CS) respiration events, and periodic breathing (PB) events were prospectively collected from three (3) independent, registered sleep technologists (RPSGT) for the N=100 Adult and N=47 Pediatric subjects, and manual scoring for sleep stage and respiratory rate events for the N=100 Respiratory Rate subjects, selected as the study population sample, by randomized, double-blinded assignment of each subject to three (3) additional, prospective scorers from the panel of N=9 total independent registered sleep technologists. The designated comparative reference was constructed using 2/3 Majority Scoring to evaluate the subject EnsoSleep device performance versus the predicate EnsoSleep device (K162627) performance for all event types and experimental endpoints evaluated, and the reference device Michele Sleep Scoring device (K112102) performance for hypopnea and central sleep apnea event types specifically. Objective performance benchmarks and acceptance criteria in terms of positive agreement (PA), negative agreement (NA), overall agreement (OA), mean absolute error (MAE), Deming Regression analysis slope and intercept coefficients, Bland-Altman mean difference and 95% upper and lower limits of agreement, each with 95% two-sided bootstrap confidence intervals (R=2000), were predefined competitively based on analysis of reported performance in the predicate device 510(k) event detection and diagnostic agreements reported. In particular, the acceptance criteria were selected based on PA, NA, OA, MAE, Deming Regression coefficient, Bland-Altman mean difference and limits of agreement performance criteria that validate performance substantially equivalent to, or greater than, but not lesser than by more than 10% or similarly defined criteria in any category of the predicate 510(k) reported device performance across all endpoints respectively.
Clinical validation testing performance analysis was collected with the final EnsoSleep software release version, final revision level, final design specification, final instructions for use, and indications for use, in the intended use environment, on the intended use population, and by the intended users, in order to validate the substantial equivalence, and the safety and effectiveness of the subject EnsoSleep device for its intended use. EnsoSleep device performance was evaluated using the defined cross-sectional experimental design, statistical methodology, and set of comprehensive experimental controls, across the following four (4) experimental endpoints:
-
- EnsoSleep is intended to assist clinicians with the assessment of sleep quality, therefore performance of device sleep scoring must be validated.
-
- EnsoSleep is intended to assist clinicians with the scoring sleep disordered breathing events used in diagnostic evaluation, therefore device performance for diagnosing sleep apnea must be validated.
-
- EnsoSleep is intended to analyze physiological signals and automatically score sleep study results, including detection of SDB events, Hypopnea events, Apnea events, including OSA events, CSA events, Arousal events, Limb Movement events, RERA events, CS events, and PB events, therefore device performance for detecting each event type must be validated.
-
- EnsoSleep is intended to analyze physiological signals and automatically score sleep study results, including detection of Respiratory Rate events, Sleep-Wake events, and Total Sleep Time, therefore device performance for detecting each event type must be validated.
{13}------------------------------------------------
The final experimental results and statistical analysis were reported for each endpoint. In total PA, NA, OA. Kappa and two-sided 95% bootstrap median confidence intervals (R=2000) were calculated by pooled-epochs versus 2/3 Majority Scoring in 20 event detection experiments evaluating 6 sleep staging events (Wake, N1, N2, N3, REM, Total) in 3 samples (adult, pediatric, and RR), 11 scoring events (SDB, HYP 1.a. (3%), HYP 1.b. (4%), Apnea, OSA, CSA, Arousal, Limb Movement, RERA, CSE, PBE) in 2 samples (adult and pediatric ), and 3 physiologic analysis events (RR. Sleep-Wake. TST) with PA. NA. OA. Kappa. Deming Regression 01 slope and B0 intercept regression coefficients, and Bland-Altman analysis, including mean differences (MD), upper limits of agreement (ULOA), and lower limits of agreement (ULOA) with twosided 95% bootstrap confidence intervals for Sleep-Wake, TST in 3 samples (adult, pediatric, and RR) as well as mean absolute error (MAE) and percent epochs ±2 breaths per minute for respiratory rate (RR) scoring in the RR adult sample, and 6 sleep apnea diagnostic severity categories (AHI ≥ 5, AHI ≥ 15, REM-AHI ≥ 15, REM-AHI ≥ 15, subgroup AHI ≥ 1, subgroup AHI ≥ 10) in 2 samples (adult and pediatric) with PA, NA, OA, Kappa, two-sided 95% bootstrap median confidence intervals (R=2000), and positive and negative likelihood ratio pairs, across all 4 experimental endpoints respectively.
The subject EnsoSleep device event detection and diagnostic agreement performance were observed to meet or exceed the PA, NA, and OA performance acceptance in the 26 total experiments (26/26) across all 4 experimental endpoints evaluated, including all 20 event detection experiments (20/20) and all 6 diagnostic agreement experiments (6/6), and with all 4 experimental endpoints (4/4) statistically analyzed and evaluated in two or more samples (2) per each endpoint respectively (2/2 Adults and Pediatrics in Endpoints 2 and 3/3 Adults, Pediatrics, and RR in Endpoints 1 and 4):
- For sleep staging events Endpoint 1, all 3 EnsoSleep PA, NA, and OA point-estimates vs ● 2/3 Majority Scoring were observed to be greater than the predicate device PA, NA, and OA point-estimates vs 2/3 Majority Scoring in some events in the Adult Sample (Wake, N2, REM, Total), Pediatric Sample (Wake, N1, N2, N3, REM, Total,), and RR Sample (Wake. N2. N3. REM. Total). Additionally, some of those event detection differences that were in all 3 performance categories (PA/NA/OA) represented a statistically significant result in terms of higher agreement with 2/3 Majority Scoring, based on low/upper-bound comparison of two-sided 95% bootstrap percentile method confidence intervals to the predicate performance, in each sample: Adult Sample (REM). Pediatric Sample (Wake, N3, Total), and RR Sample (Wake, N2, REM, Total). None of the 6 events evaluated were observed with PA, NA, or OA point-estimates vs 2/3 Majority Scoring that were 10% or lower in any of the sleep staging event types evaluated.
- For sleep apnea diagnostic agreement Endpoint 2, all 3 EnsoSleep PA, NA, and OA ● point-estimates vs 2/3 Majority Scoring were observed to be greater than the predicate device PA, NA, and OA point-estimates vs 2/3 Majority Scoring in some OSA severity categories in the Adult Sample (AHI ≥ 5) and Pediatric Sample (AHI ≥ 1). There were no samples or OSA severities for which there were statistically significant differences observed in all 3 performance measures (PA/NA/OA), based on low/upper-bound comparison of two-sided 95% bootstrap percentile method confidence intervals. With the exception of Pediatric Sample AHI > 10 PA with an observed comparison of 12.6%, none of the 6 OSA severity categories evaluated were observed with PA, NA, or OA point-
{14}------------------------------------------------
estimates vs 2/3 Majority Scoring that were 10% or lower in any of the 3 diagnostic agreement performance criteria evaluated (PA/NA/OA) respectively.
- . For scoring events Endpoint 3, all 3 EnsoSleep PA, NA, and OA point-estimates vs 2/3 Majority Scoring were observed to be greater than the designated predicate device PA. NA, and OA point-estimates vs 2/3 Majority Scoring in some events in the Adult Sample (SDB, HYP, OSA, CSA, Arousal) and the Pediatric Sample (SDB, HYP, CSA, Arousal, RERA). Additionally, some of those event detection differences that were in all 3 performance categories (PA/NA/OA) represented a statistically significant result, based on low/upper-bound comparison of two-sided 95% bootstrap percentile method confidence intervals, in each sample respectively: Adult Sample (SDB, OSA, and Arousal), and Pediatric Sample (SDB. HYP. Arousal. RERA). None of the 12 events evaluated were observed with PA, NA, or OA point-estimates vs 2/3 Majority Scoring that were 10% or lower in any of the scoring event types evaluated.
- For physiologic analysis events Endpoint 4. for SW and RR events. all three EnsoSleep . subject device PA, NA, OA, percent epochs ≤2 brpm, and/or MAE point-estimates vs 2/3 Majority Scoring were observed to be statistically similar to the predicate device PA, NA, and OA point-estimates vs 2/3 Majority Scoring in all SW and RR event types evaluated. None of the PPG-SW events evaluated were observed with PA, NA, OA pointestimates vs 2/3 Majority Scoring that were 10% or lower in any of the EEG-SW sleep staging event types evaluated in the adult, pediatric, or in the RR Samples respectively, and a global minima PPG-SW performance statistic of 89% PA in the RR Sample (4% abs diff vs 93% EEG-SW). No statistically significant differences were observed between PPG-RR and EB-RR respiratory rate events vs 2/3 Majority Scoring; by ≥90% percent epochs with RR-value within ≤2 brpm of 2/3 Majority Scoring RR-value, and by MAE of RR value within ≤ 2 brpm of 2/3 Majority Scoring RR-value, when both were evaluated and compared in the RR Sample respectively. For TST the Deming Regression coefficient parameters Slope B1 was near unity with 0.90< B1 < 1.10, Intercept B0 was near zero with B0 <15 minutes, the Bland-Altman absolute mean difference was near zero with MD within ≤15 mins, and Bland-Altman 95% absolute upper and lower limits of agreement were within <90 mins. No clinically significant deviations observed in PPG-TST or EEG-TST index values vs 2/3 Majority Scoring based the Deming regression coefficients, the average differences, the Bland-Altman 95% limits of agreement, or the two-sided 95% bootstrap confidence intervals, when compared to 2/3 Majority Scoring TST in each of the adult, pediatric, and RR samples respectively when comparing TST index performance.
- In supplemental clinical data and analyses provided with Endpoint 2, both the pediatric ● and adult samples showed a strong sleep apnea diagnostic agreement. The results showed no statistically significant difference in performance between the pediatric sample and both the predicate device and the adult sample performance which further demonstrated substantial equivalence in performance for the pediatric subgroup. The additional pediatric subgroups analyzed showed consistent results with prior pediatric study results reported in endpoints 1-4. For supplemental clinical data and analyses provided with Endpoint 4 the RR, adult, and pediatric sample groups met all criteria specified for both the Deming regression analysis and Bland-Altman analysis, with evaluated results that demonstrated strong clinical performance based on a Deming regression coefficient slope
{15}------------------------------------------------
ß 1 that was near unity and intercept β0 near zero, and all Bland-Altman limits of agreement were within 90 minutes and a mean difference within 15 minutes.
{16}------------------------------------------------
STUDY RESULTS
Sleep Staging Event Detection
The subject device EnsoSleep and predicate device EnsoSleep (K162627) clinical performance results compared for Endpoint 1, sleep staging. For all sleep staging event types evaluated in Endpoint 1, the point-estimates of the subject device PA, NA and OA event detection performance exceeded, were equivalent to, or were within 5% of the predicate device PA, NA and OA performance. The subject device PA, NA, and OA performance was observed to be statistically significantly greater in nearly all comparisons relative to the predicate device, showing a significant increase in Total Staging PA by 9% and resulting Total PA of 87% in Adult and 89% In Pediatric samples. The results confirm EnsoSleep achieves clinical performance for sleep staging positive, negative, and overall agreement that is substantially equivalent to the predicate device positive, negative, and overall agreement across all sleep stages.
| Adult Sample Pooled-Epochs EnsoSleep vs 2/3Majority Sleep Staging Performance (K210034) | Adult Sample Pooled-Epochs Predicate vs 2/3Majority Sleep Staging Performance (K162627) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| (N=100,84,408PSGepochs) | Percent Agreement (%)with two-sided 95% bootstrap medianpercentile method confidence intervals(R=2000) | Positive | Negative | Overall | (N=72,59,719PSGepochs) | Percent Agreement (%)with two-sided 95% bootstrapmedian percentile method confidenceintervals (R=1000) | Positive | Negative | |
| Overall-Epochs assigned by 2/3 Majority Scoring | Wake | 23,596 | 93.5%(93.1%,93.8%) | 97.2%(97.1%,97.4%) | 96.1%(96.0%,96.3%) | 17,459 | 86%(82%,88%) | 97%(95%,98%) | 94%(92%,95%) |
| N1 | 4,406 | 37.0%(35.6%,38.5%) | 98.3%(98.2%,98.4%) | 95.0%(94.8%,95.1%) | 3,293 | 41%(33%,48%) | 94%(93%,96%) | 91%(90%,93%) | |
| N2 | 37,890 | 88.3%(87.9%,88.6%) | 89.3%(89.0%,89.6%) | 88.8%(88.6%,89.0%) | 26,839 | 77%(73%,81%) | 87%(85%,90%) | 83%(80%,85%) | |
| N3 | 6,513 | 80.0%(79.0%,81.0%) | 96.3%(96.2%,96.5%) | 95.0%(94.9%,95.2%) | 5,587 | 81%(74%,88%) | 93%(91%,95%) | 92%(90%,94%) | |
| REM | 9,400 | 90.9%(90.4%,91.5%) | 99.3%(99.2%,99.3%) | 98.3%(98.2%,98.4%) | 6,541 | 79%(72%,84%) | 99%(98%,99%) | 96%(96%,97%) | |
| Total | 81,805 | 86.6%(86.4%, 86.9%) | 59,719 | 78%(77%, 80%) | |||||
| None | 2,603 | - | 1,432 | - |
{17}------------------------------------------------
| Pediatric Sample Pooled-Epochs EnsoSleep vs 2/3 Majority Sleep Staging Performance (K210034) | Adult Sample Pooled-Epochs Predicate vs 2/3 Majority Sleep Staging Performance (K162627) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| (N=47,38,568 PSGepochs) | Percent Agreement (%) with two-sided 95% bootstrap median percentile method confidence intervals (R=2000) | (N=72, 59,719 PSG epochs) | Percent Agreement (%) with two-sided 95% bootstrap median percentile method confidence intervals (R=1000) | ||||||
| Overall-Epochs assigned by 2/3 Majority Scoring | Overall | Total | Positive | Negative | Overall | Total | Positive | Negative | Overall |
| Wake | 7,867 | 93.1%(92.5%,93.6%) | 99.2%(99.1%,99.3%) | 97.9%(97.8%,98.1%) | 17,459 | 86%(82%,88%) | 97%(95%,98%) | 94%(92%,95%) | |
| √1 | 1,263 | 43.2%(40.4%,45.9%) | 98.8%(98.7%,99.0%) | 97.0%(96.8%,97.1%) | 3,293 | 41%(33%,48%) | 94%(93%,96%) | 91%(90%,93%) | |
| √2 | 17,542 | 92.6%(92.3%,93.0%) | 89.4%(89.0%,89.8%) | 90.9%(90.6%,91.2%) | 26,839 | 77%(73%,81%) | 87%(85%,90%) | 83%(80%,85%) | |
| √3 | 6,852 | 92.3%(91.6%,92.9%) | 97.5%(97.3%,97.7%) | 96.6%(96.4%,96.7%) | 5,587 | 81%(74%,88%) | 93%(91%,95%) | 92%(90%,94%) | |
| REM | 4,278 | 80.9%(79.6%,82.0%) | 99.1%(99.0%,99.2%) | 97.0%(96.8%,97.2%) | 6,541 | 79%(72%,84%) | 99%(98%,99%) | 96%(96%,97%) | |
| Total | 37,802 | 89.7%(89.4%, 90.0%) | 59,719 | 78%(77%, 80%) | |||||
| None | 766 | 1,432 |
{18}------------------------------------------------
Adult OSA Severity Subgroup Index
The subject device EnsoSleep and predicate device EnsoSleep (K162627) clinical performance results compared for Endpoint 2, sleep apnea diagnostic agreement. For all diagnostic agreement experiments evaluated in Endpoint 2, mild-AHI and moderate-AHI subject device PA, NA, and OA performance exceeded, were equivalent to, or were within 10% of the predicate device PA, NA, and OA performance, with statistically significant differences (greater performance) observed for NA and OA in mild-AHI. On the basis that the subject device met or exceeded objective PA, NA, and OA performance goals for sleep apnea diagnostic in all comparisons, EnsoSleep is considered substantially equivalent to the predicate device agreement for the analyzed OSA diagnostic severity groups.
| Adult Sample Per-Patient EnsoSleep and Predicate Sleep Apnea Diagnostic Agreement | ||||||||
|---|---|---|---|---|---|---|---|---|
| EnsoSleep vs 2/3 Majority Scoring AdultSample (K210034) | Predicate vs 2/3 Majority ScoringAdult Sample (K162627) | |||||||
| Enso Sleep | EnsoSleep REM | Predicate | Predicate REM | |||||
| AHI ≥ 5 | AHI ≥ 15 | AHI ≥ 5 | AHI ≥ 15 | AHI ≥ 5 | AHI ≥ 15 | AHI ≥ 5 | AHI ≥ 15 | |
| Sample size (n) | 100 | 100 | 100 | 100 | 72 | 72 | 72 | 72 |
| Positive Percent Agreement (%)with two-sided 95% bootstrapmedian percentile CI's (R=2000) | 94.4%(89.0%,98.7%) | 94.0%(85.7%,100.0%) | 86.7%(77.6%,95.0%) | 81.5%(65.0%,95.5.%) | 91%(82%,98%) | 95%(83%,100%) | 83%(72%,94%) | 79%(56%,94%) |
| Negative Percent Agreement (%)with two-sided 95% bootstrapmedian percentile CI's (R=2000) | 89.7%(75.8%,100.0%) | 96.3%(90.9%,100.0%) | 83.0%(71.1%,93.6%) | 93.3%(86.8%,98.6%) | 76%(61%,90%) | 98%(94%,100%) | 89%(75%,97%) | 96%(87%,100%) |
| Overall Percent Agreement (%)with two-sided 95% bootstrapmedian percentile CI's (R=2000) | 93.0%(88.0%,97.0%) | 95.0%(90.0%,100.0%) | 85.0%(78.0%,92.0%) | 90.0%(84.0%,95.0%) | 85%(77%,92%) | 97%(93%,100%) | 86%(76%,93%) | 92%(85%,97%) |
| Likelihood ratio (+) | 9.146(3.879,∞) | 25.458(10.154,$∞$ ) | 5.069(2.962,13.597) | 12.052(5.977,55.250) | 3.76 | 52.25 | 7.71 | 22.00 |
| Liklihood ratio (-) | 0.062(0.014,0.127) | 0.062(0.000,0.151) | 0.162(0.060,0.278) | 0.198(0.049,0.384) | 0.12 | 0.05 | 0.19 | 0.22 |
{19}------------------------------------------------
Pediatric OSA Severity Subgroup
The subject device EnsoSleep and predicate device EnsoSleep (K162627) clinical performance results compared for supplemental study results of Endpoint 2, pediatric and predicate adult sleep apnea diagnostic agreement. The supplemental pediatric subgroup statistical analysis results showed strong sleep apnea diagnostic agreement performance across a range of clinically relevant subgroup thresholds appropriate for pediatric patients. Based on review of each of the agreement performance measures and the two-sided 95% bootstrap CIs, observations confirmed no statistically significant differences in pediatric patient PA, NA, or OA performance, neither by relative nor absolute comparisons, to the predicate device adult patient PA, NA, or OA performance observed. On the basis that EnsoSleep met or exceeded objective performance goals for the additional pediation data provides further support that the device is substantially equivalent in pediatric subjects when compared to the performance of adult subjects in the predicate device.
| EnsoSleep vs 2/3 Majority Scoring Pediatric Sample | Predicate vs 2/3 Majority Scoring Adult Sample | |||||
|---|---|---|---|---|---|---|
| EnsoSleep (K210034) | Predicate (K162627) | |||||
| AHI ≥ 1 | AHI ≥ 5 | AHI ≥ 10 | AHI ≥ 15 | AHI ≥ 5 | AHI ≥ 15 | |
| Sample size (n) | 47 | 47 | 47 | 47 | 72 | 72 |
| Positive Percent Agreement (%)with two-sided 95% bootstrapmedian percentile CI's (R=2000) | 94.4%(85.3%,100.0%) | 90.5%(75.0%,100.0%) | 78.6%(45.5%,100.0%) | 85.7%(44.4%,100.0%) | 91%(82%, 98%) | 95%(83%, 100%) |
| Negative Percent Agreement (%)with two-sided 95% bootstrapmedian percentile CI's (R=2000) | 77.8%(50.0%,100.0%) | 100.0%(100.0%,100.0%) | 94.9%(86.5%,100.0%) | 100.0%(100.0%,100.0%) | 76%(61%, 90%) | 98%(94%, 100%) |
| Overall Percent Agreement (%)with two-sided 95% bootstrapmedian percentile CI's (R=2000) | 89.4%(80.9%,97.9%) | 95.7%(89.4%,100.0%) | 91.5%(83.0%,97.9%) | 97.9%(93.6%,100.0%) | 85%(77%, 92%) | 97%(93%, 100%) |
| Likelihood ratio (+) | 4.190(1.892, ∞) | ∞(∞, ∞) | 15.692(5.067, ∞) | ∞(∞,∞) | 3.76 | 52.25 |
| Liklihood ratio (-) | 0.070(0.000,0.205) | 0.095(0.000,0.250) | 0.222(0.000,0.578) | 0.143(0.000,0.556) | 0.12 | 0.05 |
{20}------------------------------------------------
Sleep Scoring Event Detection
The subject device EnsoSleep, predicate device EnsoSleep (K162627), and reference device MICHELE Sleep Scoring (K112102) clinical performance results compared for Endpoint 3, event detection agreement. For hypopneas and central apneas, the MICHELE Sleep Scoring performance testing did not calculate OA for individual event types and as such only PA and NA comparisons were established. For all event detection experiments including SDB, OSA, HYP, CSA, Arousal, and Leg Movement event types, the point-estimates of the subject device PA, NA and OA event detection performance exceeded, were equivalent to, or were within 5% of the reference device PA, NA and OA performance, with statistically significant differences (greater performance) observed in a majority of cases. On the basis that EnsoSleep met or exceeded objective PA, NA, and OA performance goals for these event types in all comparisons the device is considered substantially equivalent to the predicate device SDB, OSA, HYP, CSA, Arousal, and Leg Movement event detection performance.
| Event DetectionClinicalPerformanceComparisons | Adult Sample Pooled-Epochs EnsoSleep(K210034) vs 2/3 Majority Event DetectionPerformance | Adult Sample Overall-Pooled Predicate(K162627) and Reference Predicate (K112102)vs 2/3 Majority Event Detection Performance | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| (N=100,84,408PSGepochs) | Bootstrapped point-estimate ofmedian Percent Agreement(%) with 95% percentile bootstrapconfidence interval (R=2000resamples) | TotalEpochs | PositiveAgree-ment (PA) | NegativeAgree-ment (NA) | OverallAgree-ment (OA) | (N=72subjects,59719epochs) | Bootstrapped point-estimate ofmedian Percent Agreement(%) with 95% percentile bootstrapconfidence interval (R=1000resamples) | TotalEpochs | PositiveAgree-ment (PA) | ||
| Overall-Epochs assigned by 2/3 Majority Scoring | SleepDisorderedBreathingEvents | 8108 | 75.4%(74.5%,76.3%) | 97.0%(96.9%,97.2%) | 94.9%(94.8%,95.1%) | 4705 | 67%(58%,75%) | 93%(92%,94%) | 91%(90%,92%) | ||
| HypopneaEvents | 4420 | 66.3%(64.9%,67.6%) | 97.1%(97.0%,97.2%) | 95.5%(95.4%,95.6%) | 1822 | 60.3% | 97.6% | n/r | |||
| ObstructiveApneaEvents | 1659 | 74.1%(72.1%,76.1%) | 99.3%(99.2%,99.3%) | 98.8%(98.7%,98.8%) | 1066 | 53%(35%,71%) | 97%(96%,97%) | 96%(95%,97%) | |||
| CentralApneaEvents | 1505 | 65.3%(63.1%,67.6%) | 99.5%(99.5%,99.6%) | 98.9%(98.8%,99.0%) | 177 | 63.8% | 99.6% | n/r | |||
| ArousalEvents | 9047 | 73.6%(72.7%,74.5%) | 95.6%(95.5%,95.7%) | 93.2%(93.1%,93.4%) | 7686 | 66%(61%,71%) | 90%(88%,91%) | 87%(85%,88%) | |||
| LegMovementEvents | 6018 | 82.0%(81.0%,83.0%) | 92.4%(92.2%,92.6%) | 91.7%(91.5%,91.8%) | 5796 | 71%(60%,80%) | 90%(89%,92%) | 89%(87%,90%) |
{21}------------------------------------------------
| Event DetectionClinicalPerformanceComparisons | Pediatric Sample Pooled-Epochs EnsoSleep(K210034) vs 2/3 Majority Event DetectionPerformance | Adult Sample Overall-Pooled Predicate(K162627) and Reference Predicate (K112102)vs 2/3 Majority Event Detection Performance | |||||||
|---|---|---|---|---|---|---|---|---|---|
| (N=47,38,568PSGepochs) | Bootstrapped point-estimate ofmedian Percent Agreement(%) with 95% percentile bootstrapconfidence interval (R=2000resamples) | (N=72subjects,59719epochs) | Bootstrapped point-estimate ofmedian Percent Agreement(%) with 95% percentile bootstrapconfidence interval (R=1000resamples) | ||||||
| Overall-Epochs assigned by 2/3 Majority Scoring | TotalEpochs | PositiveAgree-ment (PA) | NegativeAgree-ment (NA) | OverallAgree-ment (OA) | TotalEpochs | PositiveAgree-ment (PA) | NegativeAgree-ment(NA) | OverallAgree-ment(OA) | |
| SleepDisorderedBreathingEvents | 1480 | 72.7%(70.4%,74.9%) | 98.6%(98.4%,98.7%) | 97.6%(97.4%,97.7%) | 4705 | 67%(58%,75%) | 93%(92%,94%) | 91%(90%,92%) | |
| HypopneaEvents | 1046 | 68.8%(66.0%,71.6%) | 98.9%(98.8%,99.0%) | 98.0%(97.9%,98.2%) | 1822 | 60.3% | 97.6% | n/r | |
| ObstructiveApneaEvents | 105 | 45.5%(36.0%,54.8%) | 99.7%(99.6%,99.7%) | 99.5%(99.5%,99.6%) | 1066 | 53%(35%,71%) | 97%(96%,97%) | 96%(95%,97%) | |
| CentralApneaEvents | 277 | 68.9%(63.5%,74.1%) | 99.7%(99.7%,99.8%) | 99.5%(99.4%,99.6%) | 177 | 63.8% | 99.6% | n/r | |
| ArousalEvents | 3018 | 78.6%(77.1%,80.0%) | 97.0%(96.8%,97.2%) | 95.5%(95.3%,95.7%) | 7686 | 66%(61%,71%) | 90%(88%,91%) | 87%(85%,88%) | |
| LegMovementEvents | 1247 | 66.0%(63.4%,68.6%) | 95.5%(95.3%,95.7%) | 94.5%(94.3%,94.8%) | 5796 | 71%(60%,80%) | 90%(89%,92%) | 89%(87%,90%) |
{22}------------------------------------------------
Total Sleep Time
The subject device EnsoSleep clinical performance results compared for supplemental study results of Endpoint 4, Total Sleep Time (TST) comparison. The reported results and observations showed that the Deming regression analysis and Bland-Altman mean difference analysis met or exceeded all acceptance criteria for each experiment type relative to the predicate device TST index agreement performance. The subject device PPG-TST and EEG-TST demonstrated statistically similar performance in all of the RR. adult, and pediatric samples respectively, demonstrated no clinically significant deviations in performance based on the Bootstrap twosided 95% confidence intervals for the observed mean differences, upper and lower limits of agreement, or regression coefficient parameters, and met or exceeded all objective performance evaluation criteria specified by the clinical study protocol and the supplemental clinical study analyses. Thus, EnsoSleep TST index agreement performance is determined to be substantially equivalent to the predicate based on acceptance criteria, and safe and effective for use.
| Per-Patient EnsoSleep Adult, Pediatric,and RR Sample Total Sleep Time IndexPerformance (K210034) | EnsoSleep PPG-TST vs 2/3Majority ScoringTST RespiratoryRate Sample | EnsoSleep EEG-TST vs 2/3Majority ScoringTST RespiratoryRate Sample | EnsoSleep EEG-TST vs 2/3Majority ScoringTST AdultSample | EnsoSleep EEG-TST vs 2/3Majority ScoringTST PediatricSample |
|---|---|---|---|---|
| Deming Regression plot Slope $β$ 1 (DRpS-$β$ 1) with two-sided 95% bootstrap medianpercentile CI's (R=2000) | 0.964(0.860, 1.067) | 0.984(0.925,1.023) | 1.037(0.974, 1.201) | 1.006(0.988, 1.018) |
| Deming Regression plot Intercept $β$ 0(DRpI- $β$ 0) [hours] with two-sided 95%bootstrap median percentile CI's(R=2000) | 0.089(-0.484, 0.663) | 0.156(-0.071,0.504) | -0.181(-1.101,0.182) | 0.021(-0.034,0.134) |
| Bland-Altman difference plot meandifference (BADp-MD) [minutes] with95% Limits of Agreement (LOAs) for 1.96 | MD: 5.380(2.372,8.475) | MD: -4.785(-6.131,-2.237) | MD: 0.515(-4.173,2.331) | MD: -3.255(-4.411,-1.472) |
| standard deviations (1.96SD) of theaverage difference including two-sided95% bootstrap-t CI's (R=2000) with the | ULOA: 73.463(68.332,78.743) | ULOA: 32.922(30.625,37.269) | ULOA: 57.750(49.751,60.849) | ULOA: 10.654(9.310,12.728) |
| MD, upper (BADp-ULOA) [minutes], andlower (BADp-LLOA) [minutes] limits ofagreement | LLOA: -62.703(-67.835,-57.423) | LLOA: -42.492(-44.789, -38.145) | LLOA: -56.720(-64.718, -53.621) | LLOA: -17.164(-18.508, -15.091 |
Clinical Performance Conclusion
Non-clinical and clinical verification, validation, and performance testing were conducted in accordance with FDA's relevant recommendations to confirm the device design met all specifications, user needs for qualified clinical and non-clinical users.
EnsoSleep has passed all of the verification and validation tests and provided clinical performance testing results that demonstrate safety, effectiveness for the intended use of the device and that the performance data demonstrate that the new technological characteristics proposed in this 510(k) submission are substantially equivalent to the predicate device.
Newly introduced sleep staging, scoring, and physiological signal-based analysis to provide event detection and sleep apnea diagnostic severity agreement performance were tested with
{23}------------------------------------------------
acceptable methods and provided data to support a determination of substantially equivalent as compared to the predicate device based on the indications for use and pre-specified acceptance criteria.
The EnsoSleep subject device performance is statistically similar to the predicate device performance, by comparison to a double-blind, prospective 2/3 Majority Scoring panel consensus reference of independent, qualified sleep technologist scorers, providing additional objective evidence that EnsoSleep is safe and effective within the indications for use. Based on the submitted clinical performance validation testing protocols, controls, methodology, evidence, and results, EnsoSleep and its sleep staging event, scoring event, physiological analysis for event detection and sleep apnea severity diagnostic agreement demonstrate substantial equivalence to the predicate device.
CONCLUSIONS
The subject device EnsoSleep and the predicate have the same intended use and are similar in basic technological characteristics to the predicate device EnsoSleep K 162627. The new device characteristics do not raise different questions of safety and effectiveness. The subject device EnsoSleep includes a new target patient population (pediatrics ages 13 and up for in-clinic and hospital PSG) and use of home sleep testing with adult patients to produce automated event detection of previously recorded PSGs.
A clinical study was performed on retrospective data of adults (18 years and older) and pediatrics (13-17 years old) and submitted in support of the new claims for the subject device. The study examined sleep scoring accuracy including hypopnea, apnea, arousal, respiratory, and movement events. One hundred adult and 47 pediatric retrospective sleep studies were scored by three clinicians and compared to the subject device automated scoring and compared to predicate and reference predicate devices. The results of the study show that the subject device automated scoring performs equal to or better than the predicate and reference devices. The clinical study confirmed the safety and effectiveness of the subject device and there were no complications or contraindications found during the study.
EnsoSleep has passed the aforementioned verification and validation tests and provided clinical performance testing results with a clinical dataset in order to demonstrate safety and effectiveness as compared to the predicate device. It is therefore concluded that the subject EnsoSleep device is substantially equivalent to the predicate EnsoSleep device.
§ 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).