Search Results

Autoplaque is intended to provide an optimized non-invasive application to analyze coronary anatomy and pathology and aid in determining treatment paths from a set of Computed Tomography (CT) Angiographic images.

Autoplaque is a workstation-based post processing application. It is a non-invasive diagnostic reading software intended for use by cardiologists and radiologists as an interactive tool for viewing and analyzing cardiac CT data for determining the presence and extent of coronary plaques and luminal stenoses.

The software is not intended to replace the skill and judgment of a qualified medical practitioner and should only be used by people who have been appropriately trained in the software's functions, capabilities and limitations. Users should be aware that certain views make use of interpolated data. This is data that is created by the software based on the original data set. Interpolated data may give the appearance of healthy tissue in situations where pathology that is near or smaller than the scanning resolution may be present.

Autoplaque must be installed on a suitable commercial computer platform. It is the user's responsibility to ensure the monitor quality and ambient light conditions are consistent with the clinical applications.

Typical users of Autoplaque are trained medical professionals, including but not limited to radiologists, clinicians, technologists, and others.

Autoplaque 3.0. a stand-alone software, performs post-processing of coronary Computed Tomography Angiography (CTA) images and measurements of components of images using computerized algorithms.

Autoplaque 3.0 aids the physician with measurement of coronary artery stenosis and provides measurements for coronary plaque and coronary artery remodeling. Autoplaque 3.0 does not replace standard clinical practice or clinician decision making.

Autoplaque 3.0 allows for standardized characterization of plaque and stenosis from DICOM image data (loaded from the local computer hard drive) and includes the following features:

• Review of heart and coronary vessels in Multiplanar Reformatting (MPR), curved MPR, and straightened MPR views:
• Measurement of vessel diameter and area;
• Characterization and measurement of plaque parameters; and
• Measurement of lumen diameter, area, and luminal stenosis.

Autoplaque 3.0 includes automated vessel, plaque and lumen segmentation, which is reviewed and can be edited, if necessary, by the clinician.

Autoplaque 3.0 can run on Windows or Mac computer platforms. The minimum hardware specifications are specified in the user manual.

Here's an analysis of the acceptance criteria and the studies conducted for Autoplaque 3.0, based on the provided text:

Acceptance Criteria and Device Performance for Autoplaque 3.0

The acceptance criteria for Autoplaque 3.0 focused on establishing concordance and agreement for various plaque and stenosis measurements compared to expert reader measurements and the predicate device. The primary statistical measures used were the intraclass correlation coefficient (ICC) and correlation coefficient. While specific numerical thresholds for "excellent agreement" or "excellent correlation" are not explicitly stated as acceptance criteria, the studies report achieving this level of agreement.

1. Table of Acceptance Criteria and Reported Device Performance

2. Sample Sizes and Data Provenance

Test Set - Comparison with Expert Reader (US Population):

• Sample Size: 201 patients, with a total of 781 lesions analyzed.
• Data Provenance: Retrospective, clinically indicated coronary computed tomography angiography (CCTA) images from multiple U.S. sites. Ethnicity information was available for 80 patients: White alone (not Hispanic or Latino) 18%, Black 19%, Hispanic or Latino 61%, and American Indian and Alaska native alone 1%.

Test Set - Comparison with Expert Reader (Non-US Population):

• Sample Size: 175 patients, with a total of 1081 lesions analyzed.
• Data Provenance: Retrospective, clinically indicated coronary computed tomography angiography (CCTA) images from multiple non-U.S. external sites.

Test Set - Comparative Study with Predicate Device:

• Sample Size: 27 patients, with a total of 30 lesions analyzed.
• Data Provenance: Not explicitly stated as retrospective or prospective, but derived from clinically indicated CCTA images obtained from a commercial Siemens CT scanner.

Training Set:

• Sample Size: Not specified in the provided text.
• Data Provenance: Not specified in the provided text.

3. Number and Qualifications of Experts for Ground Truth

For US Patient Population Study Validation:

• Number of Experts: One
• Qualifications: Board-certified imaging cardiologist.

For Non-US Patient Population Study Validation:

• Number of Experts: Two cardiologists and one additional highly experienced radiologist.
• Qualifications: Two cardiologists, one highly experienced radiologist.

4. Adjudication Method for the Test Set

For US Patient Population Study Validation:

• Adjudication Method: Not explicitly stated, as only one expert was mentioned for ground-truthing. This suggests a direct comparison method.

For Non-US Patient Population Study Validation:

• Adjudication Method: Discrepancies between the two cardiologists were resolved by an additional highly experienced radiologist. This implies a 2+1 adjudication model.

For Comparative Study with Predicate Device:

• Adjudication Method: Not applicable, as this study compared the two devices rather than devices to human ground truth.

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

No MRMC comparative effectiveness study was specifically described where human readers improved with AI vs. without AI assistance. The clinical validation studies focused on comparison of the device's output to expert measurements, rather than assessing an AI-assisted workflow for human readers. However, the significantly reduced analysis time for the device compared to experts (5.7 seconds vs. 25-30 minutes) implies a substantial efficiency improvement when using the AI tool.

6. Standalone Performance Study

Yes, a standalone performance study was conducted. Both the "Comparison with Expert Reader Measurements in a Multicenter US Patient Population" and "Comparison with Expert Reader Measurements in a Non-US Patient Population" evaluated the Autoplaque 3.0 device's performance independently against expert ground truth. The device generates its output (plaque volumes, diameter stenosis) without human intervention in the measurement process, although the output can be reviewed and edited by clinicians.

7. Type of Ground Truth Used

For comparison with expert readers (both US and Non-US populations):

• Ground Truth Type: Expert Consensus/Measurements. The ground truth was established by human expert readers (board-certified imaging cardiologist, cardiologists, and highly experienced radiologist) directly performing the measurements on the CCTA images.

8. Sample Size for the Training Set

The sample size for the training set is not explicitly mentioned in the provided text. The document focuses on the validation studies performed after the device was developed.

9. How Ground Truth for Training Set was Established

The method for establishing ground truth for the training set is not explicitly mentioned in the provided text. However, given that the device uses "Deep learning-based vessel, plaque and lumen segmentation," it is highly probable that the training data's ground truth would have been established through expert manual annotations or segmentations on CCTA images, similar to how the validation ground truth was established. The text only mentions that the deep learning-based segmentation was "validated with ground truth by clinician expert readers."

Ask a specific question about this device