Search Results

Bone VCAR is a post processing application for use in the analysis of CT images. The software is intended to support clinicians in the review of images that include the spine by providing tools to label the spine and optimize the display of anatomy within the CT image.

Bone VCAR is designed to support the clinician in visualization of the spine, by providing initial identification of vertebrae to assist in report dictation.

The software also assists the user by providing optimized display settings for easier identification of anatomy to facilitate fast image review and reporting of findings. Bone VCAR may be used for multiple care areas and is not specific to any disease state. It can be utilized during the review of exams including trauma, oncology, and routine body.

Bone VCAR is a software analysis package utilizing a deep learning technique that assists in the analysis and visualization of CT data. It is intended to provide clinicians with an optimized display and quick access to tools that improve the reading experience and efficiency for anatomy. Bone VCAR is a post processing application option for the Advantage Workstation (AW) platform, CT Scanner, Cloud or PACS stations which can be used in the analysis of CT images.

Bone VCAR is designed to support the clinician in easy visualization of spine and to provide identification of those structures to assist in report dictation. This post processing solution combines the following tools and functionality:

· Multiplanar Reconstruction (MPR) displays of axial, sagittal, coronal, oblique, x-section and curved views which can be displayed in thin/thick, Average, Maximum Intensity Projection (MIP), Minimum intensity Projection (MinIP), Volume Rendering (VR) modes

· Display of anatomical labels automatically with editing capability of all labels such as the spine.

• Display of curved views automatically and editing capability of curved views through anatomical regions such as the spine for enhanced display options

· Access to all standard volume viewer tools for measuring distances, areas, Hounsfield unit values and annotating within the images

· Synchronization of views when multiple series are loaded with spine labeling for all series loaded.

The software will assist the user by providing optimized display settings to enable fast review of the images along with easy identification of anatomy to ease reporting of findings. Bone VCAR may be used for multiple care areas and is not specific to any disease state. It can be utilized during the review of various types of exams including trauma, oncology, and routine body.

Bone VCAR is compatible with both single energy and Gemstone Spectral Imaging (GSI) acquisition methods.

Here's a summary of the acceptance criteria and study details for the Bone VCAR device, based on the provided FDA 510(k) submission:

Acceptance Criteria and Device Performance

Study Details

2. Sample Size Used for the Test Set and Data Provenance:

• Test Set Size: Not explicitly stated as a number, but referred to as "a database of exams" and "a representative set of clinical sample images."
• Data Provenance: The "database of exams is considered as a representative of the clinical scenarios where Bone VCAR is intended to be used, with consideration of acquisition parameters, image quality, pathologies and anatomy variations." No specific country of origin is mentioned. It is implied to be retrospective, as it's a "database of exams" used for validation.

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

• Number of Experts: Three.
• Qualifications: Board-certified radiologists. No specific experience length is provided.

4. Adjudication Method for the Test Set:

• Not explicitly stated. The document mentions "assessed by three board certified radiologists using 5-point Likert scale," but the process for resolving disagreements or establishing a consensus ground truth isn't detailed.

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

• Was it done? Yes, a comparative assessment was done.
• Effect Size (Improvement with AI vs. without AI assistance): The assessment "demonstrated that the capability of automatic labeling of spine by Bone VCAR is faster than manually labeling and it also increases reading and reporting efficiency whilst providing accurate identification of vertebra." An explicit effect size (e.g., percentage improvement in speed or efficiency) is not provided.

6. Standalone Performance Study (Algorithm only without human-in-the-loop):

• Was it done? Yes. The engineering validation of the algorithm's capability "provided a success rate greater than 90% for the capability of automatically labeling the spine." This indicates a standalone assessment.

7. Type of Ground Truth Used:

• For the standalone algorithm validation, the ground truth for "automatically labeling the spine" would have been established by presumably highly accurate manual labeling or anatomical expert review.
• For the human reader assessment, the "accurate identification of vertebra" suggests expert consensus or established "correct" labels against which the readers' performance (with and without AI assistance) was compared. The document doesn't specify the exact method for generating this gold standard.

8. Sample Size for the Training Set:

• Not explicitly mentioned. The document states that "Bone VCAR is a software analysis package utilizing a deep learning technique," which implies a training set was used, but its size is not provided.

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

• Not explicitly mentioned. For deep learning models, ground truth for training would typically be established through expert annotation of the images to provide the algorithm with the correct spine and vertebral labels.

Ask a specific question about this device