Search Results

Lung-CAD is a computer-assisted detection (CADe) software device that analyzes chest radiograph studies for interstitial thickening. The device uses a deep learning algorithm to identify regions of interstital thickening and produces boxes around the ROIs.

Lung-CAD is intended for use as a concurrent reading aid for physicians interpreting chest X-rays. The device is not intended for clinical diagnosis of any disease. It does not replace the role of other diagnostic testing in the standard of care for lung parenchymal findings. Lung-CAD is indicated for adults only.

Lung-CAD is computer-assisted detection (CADe) software designed to increase the accurate detection of interstitial thickening. Lung-CAD's output is available for physicians interpreting chest radiographs as a concurrent reading aid. The device helps physicians more effectively identify interstitial thickening. Lung-CAD does not replace the role of the physician or of other diagnostic testing in the standard of care and does not provide a diagnosis for any disease. Lung-CAD uses modern deep learning and computer vision techniques to analyze chest radiographs.

For each image within a study, Lung-CAD generates a DICOM Presentation State file (output overlay). If any ROI is detected by Lung-CAD in the study, the output overlay for each image includes "Interstitial thickening". 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 finding: "Interstitial thickening". If no ROI is detected by Lung-CAD in the study, the output overlay for each image will include the text "No Lung-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 Lung-CAD and a customer configurable message containing a link or instructions, for users, to access labeling documents. The Lung-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:

Acceptance Criteria and Reported Device Performance

Study Details

2. Sample Size and Data Provenance for Test Set

• Sample Size: 5,000 chest radiograph cases for the standalone performance assessment.
• Data Provenance: The text states the cases were "representative of the intended use population" but does not specify country of origin, or if they were retrospective or prospective.

3. Number of Experts and Qualifications for Ground Truth (Test Set)

• The document does not explicitly state the number of experts used to establish the ground truth for the standalone or clinical test sets, nor their specific qualifications (e.g., radiologists with X years of experience).

4. Adjudication Method (Test Set)

• The document does not explicitly state the adjudication method used for establishing the ground truth for the test set.

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

• Yes, an MRMC retrospective reader study was performed.
• Effect Size of Human Readers Improvement with AI vs. without AI assistance: The reader AUC improvement for interstitial thickening when aided by Lung-CAD versus unaided was 0.0797 (95% Confidence Interval: 0.0797, 0.0798). This improvement was statistically significant (p-value < 0.001).

6. Standalone Performance Study

• Yes, a standalone performance assessment was conducted.
• Results:
  • Sensitivity: 0.913 (95% Wilson's CI: 0.850-0.951)
  • Specificity: 0.866 (95% Wilson's CI: 0.856-0.875)
  • AUC: 0.961 (95% Bootstrap CI: 0.948-0.972)

7. Type of Ground Truth Used

• The document implies the ground truth for both standalone and clinical studies was based on expert interpretation, as it refers to "Ground Truth Positive" cases and "accuracy of readers." However, it does not explicitly state the method (e.g., expert consensus, pathology, outcome data).

8. Sample Size for Training Set

• The document does not specify the sample size used for the training set.

9. How Ground Truth for Training Set was Established

• The document does not specify how the ground truth for the training set was established. It only mentions the use of "Supervised Deep Learning," which implies that the training data was labeled.

Ask a specific question about this device