(395 days)
The TM eCloud ECG Analysis System, using proprietary algorithm, is intended for use in adults and children of any age from birth upwards. The Program makes significant use of the patient's age and gender and will provide a unique diagnosis if age differs only by a few days in the case of neonates. It is a program that is based on normal limits derived using the algorithm itself with this applying to criteria for subjects of all ages, including neonates.
TM eCloud ECG Analysis System is qualified to evaluate, detect, and aid in the physician diagnosis of the following cardiac arrhythmias and/or conduction defects:
- o Evaluation of symptoms that may be caused by cardiac arrhythmia and /or conduction disturbances
- Evaluation of symptoms that may be due to myocardial ischemia
- Detection of ECG events that alter prognosis in certain forms of heart disease
- Detection and analysis of indirect pacemaker function and failure
- Determination of cardiac response to lifestyle
- Evaluation of therapeutic interventions
- . Investigations in epidemiology and clinical trials
- Evaluation of heart rate variability in the assessment of heart disease
TM eCloud ECG Analysis System is intended to provide an interpretation of up to 12-channel ECG in all situations including resting and ambulatory ECG including Holter, cardiac event, and mobile cardiac telemetry. This software qualifies to be used in hospitals, physician offices, and scanning services. It is designed for acquisition, analysis, edit, review, report and storage of all ECG and multi-parameter data. It is capable of diagnosing all commonly recognized ECG abnormalities such as myocardial infarction (MI), acute MI, ventricular hypertrophy, abnormal ST-T changes, lethal arrhythmias, and common rhythm abnormalities.
TM eCloud ECG Analysis System can analyze recordings performed on newborns, children, and adults.
TM eCloud ECG Analysis System's interpretation results are not intended to be the sole means of diagnosis for any abnormal ECG that may be detected. It is offered to physicians and clinicians on an advisory basis only in conjunction with the physician's knowledge of ECG.
The TM eCloud ECG Analysis System consists of a (1) server-side, application Platform as a Service (PaaS) cloud based system, a (2) desktop client-side application, and (3) a web-based Physician Portal website.
The (1) server-side, application PaaS component collects, stores, performs arrhythmia analysis on ECG uploads, and transfers data to and from the client-side application.
The (2) desktop client-side application is a workstation system which allows technicians to review ECG, edit the analysis results produced by the (1) server-side, application PaaS component, and generate reports for the ECG study. The edited results and reports are uploaded to the (1) server-side, application PaaS component. It also allows notifications and updates to (3) Physician Portal website.
The TM eCloud ECG Analysis System is capable of processing and performing arrhythmia analysis on ten seconds to 60 days of recorded ECG from one to 12 channels. The system is designed to be compatible with any stationary or ambulatory ECG device having the ability to export ECG. Typical compatible devices interfaces include Cardiac Mobile Telemetry, Event recorders, and Holter recorders but is not limited to a particular ECG device. ECG interfaces from desperate devices are translated by Telemed Adapters which convert proprietary formats to the Physionet WFDB MIT format but vendors may choose to provide the Physionet WFDB directly and bypass the use of an adapter.
The purpose of the TM eCloud ECG Analysis System is to determine if any irregular rhythms, irregular beats, conduction defects, or ST depression occurred during the recording or monitoring. A qualified physician can then use the results of the analysis report to determine what action needs to be taken to help the patient reduce or prevent the occurrence of these abnormalities.
The users may upload ECG to the TM eCloud ECG Analysis System via a cellular network, REST web service, or FTP from any computer or cellular equipment. The system performs analysis on the uploaded ECG using the proprietary 12 Lead Diagnostic EKG Analysis Engine or proprietary algorithm Reduced Lead 1-7 lead Analysis Engine and downloads the results to the client.
If detected, the ECG engine will report on the following arrhythmias: ventricular ectopy, atrial ectopy, pauses, heart block, junctional rhythms, hemiblock, LBBB, RBBB, ST anomalies, and prolonged QT. It will display up to 900 statements including atrial fibrillation, ventricular fibrillation/flutter, and WPW.
The (2) client workstation software employs several screens to display the ECG data. These screens include: 5-Minute View, Hourly view, or Daily view, 8-Second View, Arrhythmias Only View, time and frequency domain HRV, ST changes, AF burden. All analyzed data is saved in a database. From there data can be organized into a report that can be printed on paper or distributed electronically.
Although the analysis engine reports arrhythmias and abnormalities with a high rate of accuracy, all results need to be reviewed and/or edited by a qualified medical professional.
The provided document is a 510(k) premarket notification summary for the TM eCloud ECG Analysis System. It outlines the device's indications for use, its description, and non-clinical tests performed to demonstrate substantial equivalence to predicate devices.
Here's an analysis of the acceptance criteria and the study that proves the device meets them, based on the provided text:
1. A table of acceptance criteria and the reported device performance:
The document does not explicitly present a table of "acceptance criteria" alongside specific reported device performance metrics in a pass/fail format. Instead, it details that testing was performed according to ANSI/AAMI EC57:2012 - Testing and reporting performance results of cardiac rhythm and ST segment measurement algorithms. This standard itself defines methodologies and expectations for performance. The summary states that "Accuracy of QRS detection analysis," "Accuracy of heart rate measurements (HRV)," "Accuracy of VEB detection analysis," "Accuracy of Ventricular Flutter or Fibrillations," "Accuracy of supraventricular ectopic beats and Atrial Flutter or Fibrillations," and "Accuracy of ST segment deviations or to detect ST changes" were tested.
However, the document does not report specific quantitative performance values (e.g., sensitivity, specificity, accuracy percentages, or limits of agreement) for any of these metrics for the TM eCloud ECG Analysis System. Therefore, a table comparing reported performance against acceptance criteria cannot be constructed from this document. The conclusion merely states that the "results show that... the software is equivalent in safety and effectiveness."
2. Sample sizes used for the test set and the data provenance (e.g. country of origin of the data, retrospective or prospective):
The testing was conducted using several publicly available datasets, which serve as the test sets for different aspects of the algorithm. The provenance and sample sizes for these databases are:
- CSE (Common Standards for Quantitative Electrocardiography) Database:
- Sample Size: 1220 ECGs
- Provenance: Recorded from individuals living in several European countries.
- Nature: Retrospective (as it's an existing database).
- Glasgow 1000 ECG Database:
- Sample Size: 1000 ECGs
- Provenance: Implied to be from Glasgow, UK.
- Nature: Retrospective. Widerange of normal and abnormal ECGs, including arrhythmias, conduction defects.
- Glasgow Adult Normal Database:
- Sample Size: 1498 ECGs
- Provenance: Implied to be from Glasgow, UK.
- Nature: Retrospective. From apparently healthy individuals examined by a physician, with no evidence of heart disease or conditions like diabetes. Used for determining normal limits of ECGs (e.g., QT interval).
- Glasgow Pediatric ECG Database:
- Sample Size: 840 ECGs
- Provenance: Implied to be from Glasgow, UK.
- Nature: Retrospective. From neonates, infants, and children referred or admitted to hospital for investigation.
- Glasgow Pacemaker ECG Database:
- Sample Size: 47 ECGs
- Provenance: Implied to be from Glasgow, UK.
- Nature: Retrospective. Selected where pacemaker stimuli were correctly detected by inspection.
- Glasgow Atrial Fibrillation Database:
- Sample Size: 72 additional cases (supplementing the 1000 ECGs database).
- Provenance: Implied to be from Glasgow, UK.
- Nature: Retrospective. Used to supplement cases of atrial fibrillation for rhythm analysis assessment.
- MIT-BIH Arrhythmia Database:
- Sample Size: Not specified directly in the document, but it's a well-known public database for arrhythmia analysis.
- Provenance: Not specified.
- Nature: Retrospective. Used for QRS detection, HR measurements, VEB detection, Ventricular Flutter/Fibrillations, supraventricular ectopic beats, Atrial Flutter/Fibrillations.
- AHA Database of Evaluation of Ventricular Arrhythmia Detectors:
- Sample Size: Not specified.
- Provenance: Not specified.
- Nature: Retrospective. Used for QRS detection, HR measurements, VEB detection, Ventricular Flutter/Fibrillations.
- Noise Stress Test Database:
- Sample Size: Not specified.
- Provenance: Not specified.
- Nature: Retrospective. Used for QRS detection, HR measurements, VEB detection, supraventricular ectopic beats, Atrial Flutter/Fibrillations.
- Congestive Heart Failure RR Interval Database:
- Sample Size: Not specified.
- Provenance: Not specified.
- Nature: Retrospective. Used for heart rate variability (HRV) analysis.
- Creighton University Ventricular Tachyarrhythmia Database (CU):
- Sample Size: Not specified.
- Provenance: Not specified.
- Nature: Retrospective. Used for Ventricular Flutter or Fibrillations.
- European ST-T Database:
- Sample Size: Not specified.
- Provenance: Not specified.
- Nature: Retrospective. Used for ST segment deviations or ST changes.
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 does not explicitly state the number or qualifications of experts used to establish ground truth for these pre-existing databases. These databases are generally compiled with established and widely accepted ground truths, often based on expert consensus, but the specific number and qualifications of those who performed the original annotations are not detailed in this submission. For example, the Glasgow Adult Normal Database indicates ECGs were taken from individuals "examined by a physician," but this pertains to patient selection, not ground truth annotation for algorithm testing.
4. Adjudication method (e.g. 2+1, 3+1, none) for the test set:
The document does not specify the adjudication method used for establishing the ground truth of the various databases. Publicly available databases often have their ground truth established by expert review, sometimes with multiple readers and an adjudication process, but this information is not provided within this 510(k) summary.
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 document describes standalone algorithm performance testing against established databases, not a multi-reader, multi-case (MRMC) comparative effectiveness study involving human readers with and without AI assistance. Therefore, there is no reported effect size of human readers improving with AI vs. without AI assistance. The device is designed to provide an interpretation on an "advisory basis only" to physicians and clinicians.
6. If a standalone (i.e. algorithm only without human-in-the-loop performance) was done:
Yes, a standalone algorithm-only performance study was done. The sections "Databases used for Non-Clinical Testing" and "Reduced lead testing was performed in accordance with ANSI/AAMI EC57/Ed.3" clearly describe the algorithm being tested against various established ECG databases for accuracy in detecting different cardiac events. The "Conclusions from Non-clinical Testing" refer to the "software" (the algorithm) being equivalent to predicate devices.
7. The type of ground truth used (expert consensus, pathology, outcomes data, etc):
The ground truth for the test sets (the various databases) is generally based on expert consensus interpretation of the ECGs, established at the time the databases were created. For example, the MIT-BIH Arrhythmia Database is renowned for its rhythm annotations made by cardiologists. The Glasgow databases would also rely on expert interpretation. The document doesn't mention pathology or outcomes data as the primary ground truth for the ECG interpretation accuracy testing.
8. The sample size for the training set:
The document does not specify the sample size for the training set used for the TM eCloud ECG Analysis System's proprietary algorithm. It only details the databases used for non-clinical testing (i.e., the test sets).
9. How the ground truth for the training set was established:
The document does not provide information on how the ground truth for the training set was established, as the training set details are not included in this summary. It only refers to a "proprietary algorithm" and "proprietary 12 Lead Diagnostic EKG Analysis Engine or proprietary algorithm Reduced Lead 1-7 lead Analysis Engine," but not their development or training methodology.
§ 870.1425 Programmable diagnostic computer.
(a)
Identification. A programmable diagnostic computer is a device that can be programmed to compute various physiologic or blood flow parameters based on the output from one or more electrodes, transducers, or measuring devices; this device includes any associated commercially supplied programs.(b)
Classification. Class II (performance standards).