CT VScore+

CT VScore+ is a software application that automatically analyzes ECG-gated, non-contrast cardiac CT images to identify and quantify calcium in the coronary arteries for patients aged 30 and older. It generates Agatston and volume calcium scores for the left main/left anterior descending (LM+LAD), right coronary artery (RCA), and left circumflex (LCX) arteries, with optional comparison to MESA and Hoff-Kondos reference databases. Users can manually edit the automated segmentation results.

CT VScore+ uses a proprietary deep learning-based Attention UNet architecture for coronary segmentation, whereas the predicate device (GE CardIQ Suite) uses unspecified deep learning methods. Both devices operate on non-contrast cardiac CT DICOM images, generate Agatston and volume scores, allow user editing of results, and support DICOM Structured Report export. The subject device is more restrictive in scope, analyzing only coronary calcium rather than broader cardiovascular anatomy and pathology.

Not stated in this summary. The document references FDA-recognized statistical methods and established reference databases (MESA, Hoff-Kondos) but does not cite specific ISO, IEC, or ASTM standards.

CT VScore+ is substantially equivalent because it performs the same intended function as the predicate (coronary calcium scoring from non-contrast cardiac CT), uses similar deep learning-based methods with validated performance meeting all acceptance criteria (ICC >0.99 for total scores, Kappa >0.95 for risk classification), and demonstrates no new safety or effectiveness concerns. The subject device includes a subset of predicate functionality with a more restrictive indication, which does not create a new intended use.