Search Results

The device is intended to capture, and report symptomatic and asymptomatic cardiac events and continuous electrocardiogram information for long-term monitoring. After wear, ECG data from compatible monitoring devices is processed and analyzed by the ZEUS System. A final report is generated on the beat-to-beat information from the entire ECG recording. For the Zio AT service, the ZEUS System supports the capture and analysis of automatically-detected arrhythmia events, as well as the analysis of uploaded patient-triggered events.

The ZEUS System is indicated for use on patients 18 years or older who may be asymptomatic or who may suffer from transient symptoms such as palpitations, shortness of breath, dizziness, light-headedness, pre-syncope, fatigue, or anxiety and patients who are asymptomatic. The reports are provided for review by the intended user to render a diagnosis based on clinical judgment and experience. It is not intended for use on critical care patients.

The Zio® Service consists of single-patient-use monitoring devices and the Zio ECG Utilization Software (ZEUS) System, the subject device of this submission, for analysis and reporting of cardiac information derived from ECG data. The ZEUS System is a software system consisting of a collection of software modules designed to store and analyze data from compatible cardiac monitoring devices to curate a report of preliminary findings intended for use by clinicians as an aid in arrhythmia diagnosis and management.

The subject ZEUS System utilizes an artificial intelligence (AI) Al-based AutoTrigger Engine (ATE) software application for processing requests originating from the Gateway Service as part of the Zio AT system enabling asymptomatic ECG triggers, and Al-based ECG Analysis Software (ECGDL) to generate the initial ECG-based cardiac information.

The output of ECG Analysis Software of the ZEUS System is used by Certified Cardiographic Technicians (CCTs) prior to publishing the cardiac information in the patient report and is not utilized directly by the prescribing clinician or patient. The reported cardiac information includes beats, ectopic runs, ECG segments, rhythms, and heart rate measurements. Recorded ECG is processed by an automated ECG analysis platform; results are quality reviewed by CCTs, findings and associated ECG are captured in a report provided to clinicians via a secure website. For the Zio® AT Patch/Gateway, the ZEUS System provides capabilities to automatically detect clinically actionable arrhythmia during the monitoring period, as well as receive baseline, scheduled, symptomatic, and asymptomatic transmissions.

The subject of this 510(k) are proposed software modifications to the ZEUS System to allow AF/AFL burden estimate reporting in the daily reports. In addition, software modifications were made to the ECGDL software of the ZEUS System to target modest performance improvements.

The ZEUS System, subject of K222389, is substantially equivalent to its predicate device (K202527), based on non-clinical performance data. Here's a summary of the available information:

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

The provided document refers to non-clinical verification and performance testing that established the device meets its design requirements. However, it does not explicitly list specific acceptance criteria thresholds (e.g., sensitivity, specificity, accuracy targets) for arrhythmia detection or AF burden estimation, nor does it provide a table of reported device performance against such specific criteria.

Instead, it states: "The nonclinical verification and performance test results established that the device meets its design requirements and intended use, that the design differences with the cleared device do not raise new questions of safety and efficacy." And, "The verification and validation testing demonstrate that the device meets all predetermined specifications."

2. Sample size used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective)

The document does not provide details on the sample size used for the test set in terms of the number of ECG recordings or patients, nor does it specify the data provenance (country of origin, retrospective/prospective). It only mentions "nonclinical verification and performance test results."

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts (e.g. radiologist with 10 years of experience)

The document states that the output of the ECG Analysis Software is "quality reviewed by CCTs" (Certified Cardiographic Technicians) before publishing the report. However, it does not specify:

• The number of CCTs or other experts involved in establishing ground truth for the test set used in the non-clinical performance evaluation.
• The specific qualifications or experience levels of these CCTs or any other experts who might have been involved in a ground truth process for testing.

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

The document does not describe any specific adjudication method (e.g., 2+1, 3+1) for the test set. It mentions CCTs perform a "quality review," but the procedure for resolving discrepancies or establishing a "ground truth" for formal testing purposes is not detailed.

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

No MRMC comparative effectiveness study is mentioned in the provided text. The document explicitly states: "No clinical testing was performed in support of this premarket notification." The system's output is intended for review by clinicians, but the document does not describe any studies on the impact of the AI on human reader performance.

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

Yes, a standalone performance of the algorithm is implied for the "ECGDL software of the ZEUS System to target modest performance improvements" and the "Al-based AutoTrigger Engine (ATE) software application." The non-clinical performance testing would evaluate the algorithm's output prior to human review, to ensure it meets design specifications. However, specific standalone performance metrics (e.g., sensitivity, specificity for various arrhythmias) are not provided in this document.

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

The document indicates that "the output of ECG Analysis Software of the ZEUS System is used by Certified Cardiographic Technicians (CCTs) prior to publishing the cardiac information in the patient report." This strongly suggests that for the operational process, the ground truth is essentially expert consensus/review by CCTs. For the non-clinical performance testing, it is implied that a similar expert review or a reference standard based on expert interpretation of ECG data would be used to establish ground truth, but this is not explicitly detailed.

8. The sample size for the training set

The document does not provide any information regarding the sample size used for the training set of the AI-based AutoTrigger Engine (ATE) and ECGDL software.

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

The document does not specify how the ground truth for the training set was established. It describes the general function of the system and its review process by CCTs, but details concerning training data annotation are absent.

Ask a specific question about this device

The ZEUS System (Zio Watch), as part of the Zio Watch Service, is intended to process and analyze electrocardiogram (ECG) and photoplethysmogram (PPG) based data to detect and report on the presence of Atrial Fibrillation (AF) over the monitoring period. The report provides ECG information for the intended user to diagnose AF and contextual information for AF, both to be interpreted based on clinical judgment and experience. It is indicated for use on adult patients 22 years or older who are susceptible to developing or who have been diagnosed with AF. It is not intended for use on critical care patients.

The ZEUS System (Zio Watch), the subject device of this 510(k) submission, is a software as a medical device (SaMD) system consisting of a collection of modules designed to process and analyze data from the Zio Watch into a curated report of preliminary findings intended for use by clinicians to aid in AF diagnosis.

The subject ZEUS System utilizes an artificial intelligence (Al) based ECG Analysis Software (ECGDL) to generate the initial ECG-based cardiac information provided to the clinician in Transmission Reports. In addition, continuously recorded PPG-based data is processed by a separate artificial intelligence (AI) based analysis software, the AF Context Engine (ACE), that detects the presence of AF. Specifically, the subject ZEUS System (Zio Watch) utilizes machine learning techniques for both the ECGDL and AF Context Engine algorithms.

These results are also presented along with the ECG data in the Zio Watch Transmission Reports. The reported cardiac information includes AF detection (including PPG-based AF summary) and heart rate measurements. The ECG-based preliminary findings in the Zio Watch Transmission Reports are quality reviewed by Certified Cardiographic Technicians (CCTs) prior to publishing. After CCT review, the report containing the preliminary findings and associated ECG are provided to clinicians via a secure website.

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

1. Table of Acceptance Criteria and Reported Device Performance:

The document doesn't explicitly state numerical acceptance criteria in a table format for a specific performance study. Instead, it focuses on demonstrating substantial equivalence to predicate devices and adherence to recognized standards. However, the core performance is related to AF detection. The acceptance criteria can be inferred as the device successfully performing its intended function of detecting and reporting AF from ECG and PPG data, with sufficient accuracy for clinical interpretation.

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

• Verily Prospective Study (ACE validation set):

  • Sample Size: Not explicitly stated as a number of patients or recordings. It describes "multi-day PPG recordings" from the Zio Watch along with "ECG-based, CCT-reviewed reference rhythm labels" obtained from simultaneously worn Zio XT Patches. Patient demographics are provided (see below).
  • Provenance: Prospective study.
  • Country of Origin: Regional Demographics (USA): Midwest: 8.0%, Mountain: 32.1%, West: 32.1%, Northeast: 14.3%, South: 13.4%.

• ZWAF Database (ECGDL validation set):

  • Sample Size: Not explicitly stated as a number of recordings. It consists of "ECG data obtained from Zio Watch as used in the Study Watch AF Detection At Home study ('Verily Prospective Study')". Patient demographics are provided (see below).
  • Provenance: This dataset is derived from the "Study Watch AF Detection At Home study" (Verily Prospective Study), which is noted as retrospective in reference to its use for the ZWAF database for ECGDL validation.
  • Country of Origin: Regional Demographics (USA) from the Verily Prospective Study (Midwest: 8.0%, Mountain: 32.1%, West: 32.1%, Northeast: 14.3%, South: 13.4%).

Demographics for both validation sets:

• Patients: At least 22 years or older who are at risk of having an AF event, as determined by having a diagnosis of paroxysmal AF.
• Age: Median = 67 [25%, 75%] = [59, 73]
• Gender: 45.5% Female

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

• Number of Experts: Not explicitly stated as a specific number. The text refers to "Certified Cardiographic Technicians (CCTs)". It implies multiple CCTs were involved in reviewing the reference ECG data to establish ground truth.
• Qualifications of Experts: "Certified Cardiographic Technicians (CCTs)". No further details on their years of experience are provided in this document.

4. Adjudication Method for the Test Set:

• The document states that the ground truth for the validation sets was based on "ECG-based, CCT-reviewed reference rhythm labels". This suggests that the CCTs' review served as the gold standard.
• No specific multi-expert adjudication method (like 2+1 or 3+1) is mentioned for resolving discrepancies among CCTs, if any. The phrasing "CCT-reviewed" implies their consensus or, at minimum, their expert determination.

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

• No MRMC comparative effectiveness study was mentioned or performed. The document explicitly states: "No clinical testing was performed in support of this premarket notification." The focus was on the performance of the algorithm itself against expertly reviewed ground truth.

6. Standalone (Algorithm Only) Performance:

• Yes, standalone performance was done for both algorithms. The validation studies described for the "AF Context Engine (ACE)" and "ECG Deep Learned (ECGDL)" algorithms assessed their performance in detecting AF and analyzing ECG data, respectively, against CCT-reviewed ground truth, without a human-in-the-loop component during the performance evaluation for substantial equivalence. The "ZEUS System (Zio Watch)" itself is described as a "software as a medical device (SaMD) system consisting of a collection of modules designed to process and analyze data".

7. Type of Ground Truth Used:

• Expert Consensus (CCT-reviewed ECG data): The primary ground truth for both the ACE and ECGDL algorithm validation was derived from "ECG-based, CCT-reviewed reference rhythm labels". For ACE, these were from simultaneously worn Zio XT Patches. For ECGDL, it was ECG data obtained from the Zio Watch itself, which was also subject to CCT review to establish ground truth.

8. Sample Size for the Training Set:

• Not explicitly stated as a number of recordings or patients. The text mentions "thousands of recordings" for both ECGDL and ACE algorithms.

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

• The training data for both the ECGDL and ACE algorithms came from "continuous cardiac recordings from compatible cardiac monitors" that "have already undergone Certified Cardiographic Technician (CCT) review." This indicates that the ground truth for the training data was also established by CCTs.

Ask a specific question about this device