Search Results

Hi-D Imaging 4TAVR is a software as a medical device (SaMD) intended to support trained healthcare professionals (cardiologists, radiologists, clinical specialists) with the assessment of the cardiovascular anatomy and visualization and measurement of structures of the heart and vessels, including the aortic valve, and with pre-operational planning and sizing for transcatheter aortic valve replacement (TAVR) procedures by providing:

• Automatic segmentation of the full heart and vessels
• 3D and 2D rendering of DICOM images
• Automatic centerline detection
• Automatic detection of the LVOT (left ventricular outflow tract), aortic, sinus, sinotubular junction (STJ), ascending aorta and descending aorta planes
• Automatic multiplanar reconstruction (MPR) based on the centerline
• Measurement and visualization tools
• Manual measurement option for the user
• Automatic S-curve extraction
• Automatic report generation

Hi-D Imaging 4TAVR software is indicated for patients with tricuspid aortic stenosis who are eligible for TAVR operations and without any prior aortic valve implants.

Hi-D Imaging 4TAVR ("the Device" or "4TAVR") is a cloud-based software as a medical device (SaMD) intended for use by healthcare professionals for the anatomy assessment of cardiac structures prior to Transcatheter Aortic Valve Replacement (TAVR) operations. 4TAVR enables healthcare professionals to maximize information using CT scan data and provide support for patient-specific pre-procedural planning of TAVR operations.

4TAVR utilizes an artificial intelligence (AI) deep learning model and computer vision methods that provide:

1. Fast and reliable automated analysis,
2. Fully automated segmentation of CT-based cardiac anatomy data,
3. 3D reconstruction and visualization of the anatomy, and
4. Fully automated report generation.

The main features of 4TAVR are:

• Fully automated extraction of heart components;
• Multiplanar reconstruction of the clinically-relevant anatomical planes for TAVR operations (i.e., left ventricular outflow tract (LVOT), annulus, sinus, sinotubular junction (STJ), ascending aorta, and descending aorta);
• 3D surface rendering of each heart component, including centerline detection;
• Fully automated detection and calculation of clinically relevant anatomical parameters such as annular perimeter, annular area, annular minimal and maximal diameter, LVOT diameter, ostium left main height and right coronary artery height, sinus of Valsalva, ascending and descending aorta perimeters, areas and diameters, sinotubular junction diameter and calcification for TAVR planning; and
• 3D downloadable meshes of aorta, left ventricle, left atrium, myocardium, right ventricle, calcium, and pulmonary artery.

4TAVR automatically processes DICOM images uploaded by the user and reports the results back to the user. The results include automatically generated measurements, 2D and 3D segmentation visualization, and a comprehensive report.

Here's an analysis of the acceptance criteria and the study that proves the device meets them, based on the provided FDA 510(k) clearance letter for Hi-D Imaging 4TAVR:

Hi-D Imaging 4TAVR is a Software as a Medical Device (SaMD) intended to support healthcare professionals in pre-operational planning and sizing for Transcatheter Aortic Valve Replacement (TAVR) procedures. The device uses AI deep learning to provide automated analysis, segmentation, 3D reconstruction/visualization, and report generation from CT scan data.

Acceptance Criteria and Reported Device Performance

Study Information

2. Sample size used for the test set and the data provenance:

• Test Set Sample Size: 929 CT scans in total across five separate studies.
• Data Provenance:
  • 3 studies from the U.S. (289 CT scans)
  • 2 studies from Europe (640 CT scans)
• Retrospective/Prospective: Not explicitly stated, but the description of comparing device measurements against manually performed ground truth measurements on existing CT scans suggests a retrospective approach.

3. Number of experts used to establish the ground truth for the test set and the qualifications of those experts:

• Number of Experts: Not explicitly stated how many distinct experts, but it mentions "experts" in plural and "the mean of observers' ground truth measurements," implying multiple experts.
• Qualifications of Experts: The experts "have relevant clinical experience and follow clinical guidelines for the measurements." No specific years of experience or titles (e.g., radiologist, cardiologist) are provided, beyond the device's intended users being cardiologists, radiologists, and clinical specialists.

4. Adjudication method for the test set:

• The ground truth was established by comparing 4TAVR results to "the mean of observers' ground truth measurements." This implies a form of consensus or average between multiple expert readers, but the precise adjudication method (e.g., how disagreements were resolved, if majority rule or a designated tie-breaker was used beyond simple averaging) is not detailed.

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:

• No MRMC comparative effectiveness study involving human readers with and without AI assistance was reported. The accuracy testing focused on the standalone performance of the AI device against expert ground truth, not on how the AI assists human readers.

6. If a standalone (i.e., algorithm only without human-in-the-loop performance) was done:

• Yes, the "Accuracy Testing Summary" explicitly describes standalone performance testing. The "accuracy tests focused on comparing the manually performed ground truth measurements by experts and the automatically generated measurements of 4TAVR." This confirms the algorithm-only performance was evaluated.

7. The type of ground truth used:

• Expert Consensus/Measurements: The ground truth was established through "manually performed ground truth measurements by experts" using "commercially available FDA-cleared and CE-Marked software," including the predicate device. The comparison was against "the mean of observers' ground truth measurements."

8. The sample size for the training set:

• Not provided in the document. The document only mentions the test set sample size (929 CT scans).

9. How the ground truth for the training set was established:

• Not provided in the document. As the training set size and details are absent, the method for establishing its ground truth is also not mentioned.

Ask a specific question about this device