Contour ProtégéAI+

Contour ProtégéAI+ is a software accessory designed to automatically contour anatomical structures in medical images (CT/MR) to assist trained medical professionals in applications like radiation therapy planning and quantitative analysis. It leverages machine-learning algorithms, specifically neural networks, and was rigorously evaluated through non-inferiority testing and external validation to ensure its safety, effectiveness, and time-saving capabilities in clinical settings.

MIM Software Inc. has introduced Contour ProtégéAI+, a software accessory designed to automatically contour anatomical structures within medical imaging data using machine-learning algorithms. This device is intended to assist trained medical professionals in the automated processing of digital medical images obtained from CT and MR modalities. Its capabilities include creating contours for various applications such as quantitative analysis, supporting adaptive therapy, transferring data to radiation therapy treatment planning systems, and archiving for patient management. The software operates as a server-based application, supporting both remote deployment via a cloud service and local installation on Windows, Mac, or Linux systems. This version incorporates one new neural network model and updates to three existing models through retraining with additional or revised contours.The device underwent comprehensive evaluation, including non-inferiority testing against a previously established reference device. Performance assessments focused on contour creation time savings and the clinical quality of the generated contours, utilizing both internal verification and external validation by medical professionals. The testing involved a diverse set of 770 images, meticulously re-segmented for training and validation. Results indicated that Contour ProtégéAI+ performed equivalently to or better than the reference device based on statistical metrics such as Dice and MDA scores, and user evaluations confirmed comparable time-saving functionality. While some limitations related to image quality, contrast, and patient demographics in the training data were noted, the overall testing confirmed that the device is safe and effective for clinical use, adhering to established design control and software development processes.

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 performance of Contour ProtégéAI+ vary across different patient populations or disease states not explicitly covered in the training data?; What specific types of image artifacts or low-quality imaging most significantly impact the accuracy and reliability of the AI-generated contours?; Can the software be trained or adapted to contour novel anatomical structures or pathologies not included in its current models, and what is the process for such an update?; How does the time-saving functionality translate into real-world clinical workflow efficiency, especially in high-volume departments or complex adaptive therapy scenarios?; What are the long-term data management and ethical considerations for archiving patient-specific contours generated by AI, particularly concerning privacy and model drift?