Search Results

EchoGo Heart Failure 1.0 is an automated machine learning-based decision support system, indicated as a diagnostic aid for patients undergoing routine functional cardiovascular assessment using echocardiography. When utilised by an interpreting clinician, this device provides information that may be useful in detecting heart failure with preserved ejection fraction (HFpEF).

EchoGo Heart Failure 1.0 is indicated in adult populations over 25 years of age. Patient management decisions should not be made solely on the results of the EchoGo Heart Failure 1.0 analysis.

EchoGo Heart Failure 1.0 takes as input an apical 4-chamber view of the heart that has been captured and assessed to have an ejection fraction ≥50%.

EchoGo Heart Failure 1.0 is a software-only medical device manufactured by Ultromics Limited and granted breakthrough status by the FDA under Q212613.

EchoGo Heart Failure 1.0 takes as input a DICOM file containing an echocardiogram and reports a classification decision suggestive of the presence or absence of heart failure with preserved ejection fraction (HFpEF). The output of this device is based on an artificial intelligence (Al) model developed using a convolutional neural network that produces the classification result. The model takes as input a 2D echocardiogram in which an apical 4-chamber view of the heart has been captured and assessed to have an ejection fraction ≥50% (this would normally be computed using a medical device for the assessment of cardiac function of the left ventricle, for example K213275). The echocardiogram should be acquired without contrast and contain at least one full cardiac cycle.

Independent training, validation and test datasets were used for training and performance assessment of the device. EchoGo Heart Failure 1.0 is fully automated and does not comprise a user interface.

EchoGo Heart Failure 1.0 produces a report containing the result of the classification, and this report is intended to be used by an interpreting clinician as an aid to diagnosis for HFpEF. The results are intended as an additional input to standard diagnostic pathways and should only be used by an interpreting clinician. The device is a diagnostic aid and thus according to common medical sense and the principles of differential diagnosis any diagnostic finding derived from usage of this product must be confirmed by additional diagnostic investigations, if in doubt. The ultimate diagnostic decision remains the responsibility of the interpreting clinician using patient presentation, medical history, and the results of available diagnostic tests, one of which may be EchoGo Heart Failure 1.0.

EchoGo Heart Failure 1.0 is a prescription only device.

Here's a breakdown of the acceptance criteria and the study proving the device meets those criteria, based on the provided text:

1. Table of Acceptance Criteria and Reported Device Performance

The provided text only explicitly states the reported device performance and the p-values for one-sided binomial exact tests against a priori acceptance criteria. The specific numerical acceptance criteria (e.g., minimum sensitivity and specificity thresholds) are not directly stated in the document. However, since the p-values are reported as p<0.001, it implies that the reported performance met or exceeded the pre-specified thresholds.

Note on Acceptance Criteria: The document mentions "a priori acceptance criteria" for sensitivity and specificity but does not specify their numerical values. The p-values of <0.001 indicate that the observed sensitivity and specificity are statistically significantly higher than these unstated thresholds.

2. Sample Size for the Test Set and Data Provenance

• Sample Size for the Test Set: 1,285 patients (comprising 639 controls and 646 cases).
• Data Provenance: Retrospective case-control study. The data was collected from multiple independent clinical sites spanning five states in the USA.

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

The document does not explicitly state the number of experts used to establish the ground truth for the test set or their specific qualifications (e.g., "radiologist with 10 years of experience"). It mentions that the "device output was compared to the ground truth classifications of cases (HFpEF) or controls."
We can infer that the ground truth for human-in-the-loop performance would be established in conjunction with "accredited cardiac physiologists (N=2) and cardiologists (N=5)" or similar, as cited in Section 7 for Usability for the formative and summative evaluations. However, this is for usability assessment and not explicitly tied to the ground truth creation for the performance study.

4. Adjudication Method for the Test Set

The document does not specify the adjudication method used for the test set.

5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study Was Done

The document does not indicate that a Multi-Reader Multi-Case (MRMC) comparative effectiveness study was done. Therefore, no effect size of human readers improving with AI vs. without AI assistance is reported. The study described focuses on the standalone performance of the AI device against a ground truth.

6. If a Standalone (i.e., algorithm only without human-in-the-loop performance) Was Done

Yes, a standalone performance study was done. The device's output (classification) was directly compared to the ground truth classifications without human intervention in the device's output decision. The text explicitly states, "EchoGo Heart Failure 1.0 is fully automated and does not comprise a user interface" (Page 4), which further supports this.

7. The Type of Ground Truth Used

The ground truth used was "ground truth classifications of cases (HFpEF) or controls," implying a clinical diagnosis or consensus based on established diagnostic criteria for Heart Failure with Preserved Ejection Fraction. The document does not specify if it was solely expert consensus, pathology, or outcomes data, but it refers to "ground truth classifications," which typically involves expert review of comprehensive patient data.

8. The Sample Size for the Training Set

The document states, "Independent training, validation and test datasets were used for training and performance assessment of the device" (Page 4) but does not specify the sample size for the training set.

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

The document does not explicitly state how the ground truth for the training set was established. It only mentions that the AI model was "developed using a convolutional neural network" and used "independent training, validation and test datasets." It can be inferred that a similar method to the test set (clinical diagnosis/consensus) would have been used for the training data ground truth.

Ask a specific question about this device