AI-Rad Companion (Cardiovascular)

AI-Rad Companion (Cardiovascular) VA20 is an advanced software utilizing machine and deep learning algorithms to provide quantitative and qualitative analysis of cardiovascular structures from CT DICOM images, aiming to improve diagnostic support. This version enhances aorta segmentation and diameter measurement capabilities, including new maximum diameter assessments, and introduces flexible reporting and deployment options.

The AI-Rad Companion (Cardiovascular) SW version VA20 is an enhanced image processing software designed to assist clinicians in evaluating cardiovascular diseases using previously acquired Computed Tomography (CT) DICOM images from adult patients. Its core functionalities include AI-based segmentation and volume measurement of the heart, quantification of total coronary artery calcium, segmentation of the aorta, measurement of maximum aortic diameters at typical landmarks, and threshold-based highlighting of enlarged diameters. This version introduces several key improvements over its predicate device (K183268), including enhanced aorta segmentation performance through additional training data, new capabilities to measure and visualize maximum diameters of the ascending and descending aorta, individual confirmation options for aorta findings, and the generation of DICOM Structured Reports (TID 1500). Additionally, it supports both cloud and on-premise 'edge' deployment. Non-clinical software validation demonstrated improved or equivalent performance. The aorta segmentation achieved an average DICE coefficient of 0.924, up from 0.910 in the predicate. Aortic diameter measurements at nine predefined locations showed a bias within ±1.5 mm and mean absolute error (MAE) ≤2.2 mm. For the newly added maximum ascending and descending aorta diameter measurements, 91.9% of device readings fell within the ±3.51 mm inter-reader limits of agreement, with a bias within ±1.5 mm and MAE ≤1.8 mm. The software has been validated for non-cardiac chest CT data from Siemens Healthineers, GE Healthcare, Philips, and Toshiba/Canon scanners, with the calcium detection feature also validated on iterative reconstruction data from Siemens Healthineers. No new clinical tests were conducted for these specific modifications; validation relied on existing clinical data and software performance testing.

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 the aorta segmentation algorithm vary with different CT scanner manufacturers and reconstruction kernels beyond those explicitly validated?; What is the clinical impact of the new maximum ascending and descending aorta diameter measurements on patient management compared to the previous 9-landmark AHA guideline measurements alone?; Can the AI-Rad Companion (Cardiovascular) differentiate between true aortic ectasia/aneurysm and post-stenotic dilatation or other anatomical variants that might affect diameter measurements?; What is the real-world time-saving and diagnostic accuracy improvement observed by clinicians using the enhanced AI-Rad Companion (Cardiovascular) VA20 compared to manual measurements or the previous version?; How does the 'edge deployment' model ensure data security and patient privacy within the customer environment while maintaining connectivity for remote monitoring and maintenance?