Search Results

Aorta-CAD is a computer-assisted detection (CADe) software device that analyzes chest radiograph studies for suspicious regions of interest (ROIs). The device uses a deep learning algorithm to identify ROIs and produces boxes around the ROls. The boxes are labeled with one of the following radiographic findings: Aortic calcification or Dilated aorta.

Aorta-CAD is intended for use as a concurrent reading aid for physicians looking for ROIs with radiographic findings suggestive of Aortic Atherosclerosis or Aortic Ectasia. It does not replace the role of the physician or of other diagnosic testing in the standard of care. Aorta-CAD is indicated for adults only.

Aorta-CAD is computer-assisted detection (CADe) software designed for physicians to increase the accurate detection of findings on chest radiographs that are suggestive of chronic conditions in the aorta. The ROIs are labeled with one of the following radiographic findings: Aortic calcification or Dilated aorta. Aorta-CAD is intended for use as a concurrent reading aid for physicians looking for suspicious ROIs with radiographic findings suggestive of Aortic Atherosclerosis or Aortic Ectasia. Aorta-CAD's output is available for physicians as a concurrent reading aid and does not replace the role of the physician or of other diagnostic testing in the standard of care for the distinct conditions. Aorta-CAD uses modern deep learning and computer vision techniques to analyze chest radiographs.

For each image within a study, Aorta-CAD generates a DICOM Presentation State file (output overlay). If any ROI is detected by Aorta-CAD in the study, the output overlay for each image includes which radiographic finding(s) were identified and what chronic condition in the aorta is suggested by these findings, such as "Aortic calcification suggestive of Aortic Atherosclerosis." In addition, if ROI(s) are detected in an image, bounding boxes surrounding each detected ROI are included in the output overlay for that image and are labeled with the radiographic findings, such as "Aortic calcification". If no ROI is detected by Aorta-CAD in the study, the output overlay for each image will include the text "No Aorta-CAD ROI(s)" and no bounding boxes will be included. Regardless of whether an ROI is detected, the overlay includes text identifying the X-ray study as analyzed by Aorta-CAD and a customer configurable message containing a link to our instructions for users to access labeling documents. The Aorta-CAD overlay can be toggled on or off by the physician within their Picture Archiving and Communication System (PACS) viewer, allowing for concurrent review of the X-ray study.

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

1. Table of Acceptance Criteria and Reported Device Performance

The document doesn't explicitly list "acceptance criteria" in a separate section with pass/fail thresholds. Instead, it describes "performance assessment" and "clinical study" results which implicitly serve as the demonstration that the device performs acceptably and is substantially equivalent. The key performance metrics are presented from the standalone and clinical studies.

2. Sample Size for the Test Set and Data Provenance

• Standalone Test Set Sample Size: 5,000 chest radiograph cases.
• Clinical (MRMC) Test Set Sample Size: 244 cases (each reader evaluated these under both aided and unaided conditions).
• Data Provenance: The document states "cases representative of the intended use population" for the standalone study. For the MRMC study, it also doesn't explicitly state the country of origin but implies a general patient population suitable for chest radiographs. Both studies are explicitly stated as retrospective.

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

• Number of Experts: A "panel" of U.S. board-certified radiologists. The exact number is not specified beyond "a panel."
• Qualifications of Experts: U.S. board-certified radiologists. No specific experience length (e.g., 10 years) is mentioned.

4. Adjudication Method for the Test Set Ground Truth

• The ground truth for the MRMC study cases was established by "a panel of U.S. board-certified radiologists who assigned a ground truth binary label indicating the presence of Aortic calcification suggestive of Aortic Atherosclerosis and Dilated aorta suggestive of Aortic Ectasia."
• This suggests a consensus or majority vote approach, but the specific adjudication method (e.g., 2+1, 3+1, etc.) is not explicitly detailed. It only mentions "a binary label."

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

• Was an MRMC study done? Yes, a "fully-crossed multiple reader, multiple case (MRMC) retrospective reader study" was conducted.
• Effect Size of Human Reader Improvement with AI vs. Without AI Assistance:
  • Reader AUC improvement for Aortic calcification suggestive of Aortic Atherosclerosis: 0.0984 (95% Confidence Interval: 0.0984, 0.0985).
  • Reader AUC improvement for Dilated aorta suggestive of Aortic Ectasia: 0.0885 (95% Confidence Interval: 0.0885, 0.0886).
  • The study demonstrated that "Reader AUC estimates significantly improved for both categories (p-values < 0.001)."

6. Standalone (Algorithm Only) Performance Study

• Was a standalone study done? Yes. "Imagen conducted a standalone performance assessment on 5,000 chest radiograph cases..."
• Performance:
  • Sensitivity: 0.910 (0.896, 0.922)
  • Specificity: 0.896 (0.889, 0.902)
  • Overall AUC: 0.974 (0.971, 0.977)
  • Category-specific AUCs:
    • Aortic calcification: 0.972 (0.967, 0.976)
    • Dilated aorta: 0.948 (0.939, 0.957)
  • Category-specific Sensitivity/Specificity (Table 3 on page 9)

7. Type of Ground Truth Used

• The ground truth for both the standalone and MRMC studies was established by expert consensus (a "panel of U.S. board-certified radiologists"). They assigned a "binary label" (presence/absence). There is no mention of pathology or outcomes data being used for ground truth.

8. Sample Size for the Training Set

• The document does not provide the sample size used for the training set. It only mentions that the device uses "modern deep learning and computer vision techniques" and "Supervised Deep Learning."

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 only states that the device uses "Supervised Deep Learning," which implies a labeled dataset, but the method of labeling is not described.

Ask a specific question about this device