(174 days)
The CardioSTAT ECG Test Solution is intended to capture, and report symptomatic and/or continuous ECG information for long-term ambulatory monitoring (up to 14 days). It is indicated for use on adult patients who may be asymptomatic or who may suffer from transient symptoms, such as palpitations, shortness, lightheadedness, presyncope, syncope, fatigue, or anxiety. The reported ECG metrics include single-lead analysis on a beat by beat basis, heart rate measurement and thythm annotation. The report does not contain diagnostic interpretation; the report is provided for review to the intended user to render a diagnosis based on clinical judgment and experience.
The CardioSTAT® ECG Test Solution is a complete by prescription-only system for continuous ambulatory collection of ECG single-channel data for up to 14 days, followed by analysis and reporting of collected data. The CardioSTAT® ECG Test Solution is composed of two major components: (1) CardioSTAT® ECG Recorder and (2) ECG Analysis and Reporting Tool (EART) .
The CardioSTAT® ECG Recorder (1) is a nonsterile, non-invasive, and single-patientuse recorder worn on the patient's torso that provides a continuous, single-channel, ECG recording for up to 14 days. The Recorder snaps onto two off-the-shelf FDAcleared long-term monitoring electrodes on which OFIX® adhesive tape is applied to ensure long-term adherence to the patient. After the prescribed recording duration, the device is taken off by the patient and mailed to the data analysis center for analysis where data is analyzed by Certified Cardiographic Technicians (CCT) using (2) ECG Analysis and Reporting Tool (EART). The EART is intended for use only by trained technicians. A clinical report is then sent to the physician. The report does not contain diagnostic interpretation; the reported analysis is provided for review by the physician to render a diagnosis based on clinical judgment and experience.
A home use kit (CardioSTAT® Home Application) is also offered which includes all required devices to allow for remote prescription and home application of the CardioSTAT® recorder. The included devices (surgical razor, alcohol prep pad, electrodes, OFIX® tape) outside of this submission are currently available on the market.
The CardioSTAT® device is not designed to record ECG signals suitable to identify ST segment shifts and the EART software is not designed to detect or measure ST segment shifts.
The ECG signal collected by the CardioSTAT® ECG Test Solution is not intended for automated analysis other than beat and noise detection.
The ECG signal collected by the CardioSTAT® ECG Test Solution is not intended for third party ECG analysis systems.
The CardioSTAT® recorder is not a real-time device and all clinical data is reviewed after the device has been returned to Icentia.
The CardioSTAT® ECG Test Solution software provides automated detection of beats and noise only. All other classifications, rhythms and anomalies must be annotated by hand by the user in the EART software. Detections of beats and noise must be verified and confirmed by the user, or manually corrected with their respective annotation tools.
The provided FDA 510(k) summary for the CardioSTAT® ECG Test Solution does not directly present specific acceptance criteria and detailed performance data in a tabular format that is typical for a clinical study comparing an AI device against a defined ground truth. Instead, it relies on substantial equivalence to predicate devices and references to existing literature and post-market data to support its performance claims.
However, based on the information provided, we can extract details related to performance and study references.
1. Table of Acceptance Criteria and Reported Device Performance
Formal acceptance criteria for AI algorithm performance (e.g., sensitivity, specificity, accuracy against a gold standard) are not explicitly stated in the document. The performance is primarily demonstrated through comparison with "standard Holter monitoring" and the predicate Zio patch system in referenced literature.
Below is an attempt to construct a table based on the provided information, noting the lack of explicit acceptance criteria and direct performance metrics for the device's automated analysis beyond beat and noise detection.
| Criteria (Implicit) | Reported Device Performance (from literature/post-market data) | Notes |
|---|---|---|
| Arrhythmia Detection Accuracy (AF) | 99% agreement with standard Holter monitoring for AF detection | Based on a comparative study (Nault et al., 2019) with 212 patient recordings. This is the closest to a performance metric provided, specifically for Atrial Fibrillation (AF) detection. However, it's an agreement rate, not a direct sensitivity/specificity against a ground truth from an expert panel. |
| Patient Wear Time for Long-Term Monitoring | 13.1 days (SD 1 day, median 13.9 days) when prescribed for 14 days | Based on Icentia Post Market data from Canada and UK (33486 ECG tests). This is compared favorably to the Zio patch's unweighted mean wear time of 10.4 days (median 5-14 days) from a literature review (Yennikomshian et al.). |
| Signal Acquisition Quality (Vertical vs. Standard Position) | Similar data quality and patient compliance in home-based self-installation (VSL position) compared to conventional in-clinic models. | Based on a study (Benjamin et al., 2021) comparing 47 devices from each group (home vs. in-clinic). This supports the flexibility of application. |
2. Sample Size for the Test Set and Data Provenance
- Arrhythmia Detection (AF):
- Sample Size: 212 patient recordings (Nault et al., 2019).
- Data Provenance: Not explicitly stated, but the authors are from Quebec, suggesting it may be Canadian data. The study is retrospective.
- Signal Acquisition Quality (VSL position/Home Application):
- Sample Size: 47 devices in the home-based VSL group, 47 devices in the in-clinic standard application group (Benjamin et al., 2021). Total of 94 devices.
- Data Provenance: Not explicitly stated, but the authors are listed with affiliations in Israel and Canada. The study appears to be prospective due to its design for a pandemic response.
- Patient Wear Time:
- Sample Size: 33,486 ECG tests (Icentia Post Market data).
- Data Provenance: Canada and UK (retrospective post-market data).
3. Number of Experts and Qualifications for Ground Truth
The document does not explicitly state the number or qualifications of experts used to establish ground truth for the referenced studies.
- For the arrhythmia detection study (Nault et al., 2019), the "standard Holter monitoring" would have been interpreted by cardiologists or trained technicians. However, the specific number of experts, their qualifications, or the adjudication method are not detailed in this submission.
- For the signal acquisition quality study (Benjamin et al., 2021), the assessment of data quality would likely involve expert review, but this is not specified.
- The patient wear time study is based on logistical and operational data, not diagnostic accuracy that typically requires expert ground truth.
4. Adjudication Method for the Test Set
The document does not describe any specific adjudication methods (e.g., 2+1, 3+1) for establishing ground truth in the referenced studies. The Nault et al. study compares the CardioSTAT® against "standard Holter monitoring," implying the Holter interpretation served as the reference, but details of its establishment are absent.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
No formal MRMC comparative effectiveness study is mentioned in the provided text where human readers improve with AI vs. without AI assistance. The document explicitly states: "The CardioSTAT® ECG Test Solution software provides automated detection of beats and noise only. All other classifications, rhythms and anomalies must be annotated by hand by the user in the EART software. Detections of beats and noise must be verified and confirmed by the user, or manually corrected with their respective annotation tools." This indicates that the device's primary function is data capture and preliminary annotation, with significant human-in-the-loop for final interpretation.
6. Standalone (Algorithm Only) Performance
The device is not presented as a standalone diagnostic device. Its software's automated analysis is limited to "beat and noise detection only." All other classifications, rhythms, and anomalies require manual annotation and verification by trained technicians (Certified Cardiographic Technicians – CCT) using the ECG Analysis and Reporting Tool (EART). The final report is then sent to a physician for diagnosis based on clinical judgment and experience. Therefore, a standalone performance study for comprehensive arrhythmia detection by the algorithm alone (without human input for "all other classifications") is not performed or reported.
7. Type of Ground Truth Used
- For arrhythmia detection (Nault et al.), the ground truth appears to be based on the findings from "standard Holter monitoring." This would typically imply expert interpretation of Holter ECGs.
- For signal acquisition quality (Benjamin et al.), data quality assessment likely uses expert review, but the specific type of ground truth is not detailed.
- For patient wear time, the ground truth is objective data from the device/system itself (duration of recording).
8. Sample Size for the Training Set
The document does not provide information regarding the sample size used for training any internal algorithms (e.g., beat and noise detection) within the CardioSTAT® ECG Test Solution.
9. How the Ground Truth for the Training Set Was Established
The document does not provide information on how ground truth was established for any training data, as details about internal algorithm training are not disclosed.
§ 870.2800 Medical magnetic tape recorder.
(a)
Identification. A medical magnetic tape recorder is a device used to record and play back signals from, for example, physiological amplifiers, signal conditioners, or computers.(b)
Classification. Class II (performance standards).