(136 days)
VBrain is a software device intended to assist trained medical professionals, during their clinical workflows of radiation therapy treatment planning, by providing initial object contours of known (diagnosed) brain tumors (i.e., region of interest, ROI) on axial T1 contrast-enhanced brain MRI images.
VBrain uses an artificial intelligence algorithm (i.e., deep learning neural networks) to contour (segment) brain tumor on MRI images for trained medical professionals' attention, which is meant for informational purposes only and not intended for replacing their current standard practice of manual contouring process. VBrain does not alter the original MRI image, nor does it intend to be used to detect tumors for diagnosis. VBrain is intended only for generating Gross Tumor Volume (GTV) contours of brain metastases, meningiomas, and acoustic neuromas on axial T1 contrast-enhanced MRI images; It is not intended to be used with images of other brain tumors. The user must know the tumor type when they use VBrain. VBrain is intended to be used on adult patients only.
Medical professionals must finalize (confirm or modify) the contours generated by VBrain, as necessary, using an external platform available at the facility that supports DICOM-RT viewing/editing functions, such as image visualization software and treatment planning system.
VBrain is a software device indicated for use in the analysis of brain MRI images. The device consists of image preprocessing, deep learning neural networks, and postprocessing components, and is intended to assist trained medical professionals, during clinical workflows of radiation therapy treatment planning, by highlighting and contouring known (diagnosed) brain tumors on the axial T1 contrast-enhanced MRI images. The software is configured to work on a PACS network. Upon user's request, it will patient scans or users can send corresponding MR images, and the device will utilize deep learning neural networks to generate contours for the detected/diagnosed brain tumors and export the results as DICOM-RT objects (using the RT Structure Set ROI Contour attribute, RTSTRUCT) back to the network. The medical professionals must finalize (confirm and modify) the contours produced by VBrain as necessary using an external platform that supports RT DICOM viewing/editing, such as a treatment planning system.
The provided text describes the performance data for Vysioneer's VBrain device. Here's a breakdown of the acceptance criteria and the study that proves the device meets them:
1. Table of Acceptance Criteria and Reported Device Performance
| Performance Metric | Acceptance Criteria/Performance Goal (Implicitly "As Demonstrated") | Reported Device Performance (95% Confidence Interval) |
|---|---|---|
| Lesion-wise Sensitivity | Meets performance goals | 90.3% (86.1-93.7%) |
| False-Positive Rate (tumors/case) | Meets performance goals | 0.681 (0.500-0.879) |
| Lesion-wise Dice Similarity Coefficient (DSC) | Meets performance goals | 0.793 (0.775-0.811) |
| Average Hausdorff Distance (in terms of lesion size) | Meets performance goals | 5.0% (4.4-5.6%) |
| Centroid Distance (in terms of lesion size) | Meets performance goals | 5.6% (5.0-6.2%) |
Note: The document explicitly states "VBrain meets all performance goals" and "All the metrics were demonstrated to pass the performance goals," implying that the reported performance values themselves serve as the acceptance criteria being met.
2. Sample Size Used for the Test Set and Data Provenance
- Sample Size: 116 cases with 238 tumors.
- Data Provenance: Retrospective, multicenter, multinational. The data was acquired from 4 different institutions: 3 from the U.S. and 1 non-U.S.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Their Qualifications
- Number of Experts: Three.
- Qualifications of Experts: Board-certified radiation oncologists.
4. Adjudication Method for the Test Set
- Method: Consensus. The ground truth of each tumor contour was generated from the consensus of the three board-certified radiation oncologists.
5. Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study
- The document does not mention a multi-reader multi-case (MRMC) comparative effectiveness study to evaluate how human readers improve with AI vs. without AI assistance. The study focuses on the standalone performance of the VBrain algorithm relative to ground truth established by expert consensus.
6. Standalone (Algorithm Only Without Human-in-the-Loop Performance) Study
- Yes, a standalone performance study was conducted. The reported metrics (Sensitivity, False-Positive Rate, DSC, Hausdorff Distance, Centroid Distance) directly evaluate the VBrain algorithm's performance in segmenting tumors against an expert-defined ground truth, without measuring human-in-the-loop performance improvement.
7. Type of Ground Truth Used
- Type: Expert Consensus. The ground truth for tumor contours was established by the consensus of three board-certified radiation oncologists.
8. Sample Size for the Training Set
- The document does not explicitly state the sample size used for the training set. It mentions that VBrain uses an "artificial intelligence algorithm (i.e., deep learning neural networks)" which implies a training phase, but the details of the training data are not provided in this specific excerpt.
9. How the Ground Truth for the Training Set Was Established
- The document does not explicitly state how the ground truth for the training set was established. While it describes the ground truth process for the test set, it does not detail the methodology for the training data used to develop the deep learning model.
§ 892.2050 Medical image management and processing system.
(a)
Identification. A medical image management and processing system is a device that provides one or more capabilities relating to the review and digital processing of medical images for the purposes of interpretation by a trained practitioner of disease detection, diagnosis, or patient management. The software components may provide advanced or complex image processing functions for image manipulation, enhancement, or quantification that are intended for use in the interpretation and analysis of medical images. Advanced image manipulation functions may include image segmentation, multimodality image registration, or 3D visualization. Complex quantitative functions may include semi-automated measurements or time-series measurements.(b)
Classification. Class II (special controls; voluntary standards—Digital Imaging and Communications in Medicine (DICOM) Std., Joint Photographic Experts Group (JPEG) Std., Society of Motion Picture and Television Engineers (SMPTE) Test Pattern).