Contour+ (MVision AI Segmentation)

Contour+ is a medical software system designed to significantly accelerate region of interest delineation for radiation therapy treatment planning. It achieves this by employing machine learning-based algorithms to automatically generate initial segmentation templates from CT and MR images, thereby reducing the upfront effort and time required for image contouring.

The U.S. Food & Drug Administration has determined that the medical software system, Contour+ (MVision AI Segmentation), is substantially equivalent for its stated indications for use. This Class II device, regulated under 21 CFR 892.2050 for medical image management and processing systems, has met the necessary criteria. MVision AI Oy, the manufacturer, is required to comply with various regulatory provisions, including annual registration, device listing, adherence to good manufacturing practices, proper labeling, and medical device reporting. Contour+ (MVision AI Segmentation) is a software-only device designed to accelerate region of interest delineation in radiation therapy treatment planning. It utilizes machine learning-based algorithms to automatically contour CT and MR images, generating initial segmentation templates for predefined structures across multiple anatomical sites, including the Head and Neck, Brain, Breast, Lung and Abdomen, and Male and Female Pelvis. These software-generated contours are intended to be an initial template for medical professionals, requiring visualization, review, modification, and approval prior to further clinical use. The system explicitly does not detect lesions or tumors and is not intended for use with real-time adaptive planning workflows. Comprehensive verification and validation testing confirmed the system's performance, demonstrating its effectiveness in reducing the upfront effort and time required for image contouring.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the system's accuracy compare to manual contouring by expert clinicians across different anatomical sites and pathologies?; What specific machine learning algorithms are employed, and what datasets were used for training and validation to ensure robustness across diverse patient populations?; Could the software's templates lead to cognitive biases or over-reliance, potentially reducing the critical review needed by medical professionals?; Are there plans to extend the capabilities of Contour+ to include real-time adaptive planning workflows or integrate lesion detection in the future?; What is the demonstrated long-term impact of using Contour+ on overall treatment planning efficiency and patient outcomes in a clinical setting?