HealthCCSng

HealthCCSng is an AI-powered software designed to opportunistically analyze non-cardiac-gated CT scans for the detection and categorization of coronary artery calcium, providing radiologists with an estimated calcium detection category to aid in identifying potential coronary artery disease risk. Its methodology relies on automated image analysis of existing CT data to identify calcifications above 130 HU, validated through retrospective studies demonstrating high agreement with radiologist-defined ground truth, thereby offering a supplementary tool within the diagnostic workflow.

HealthCCSng by Zebra Medical Vision Ltd. is a non-invasive, post-processing software designed to aid in the evaluation of calcified plaques in the coronary arteries, which are indicative of coronary artery disease risk. It functions by analyzing existing non-cardiac-gated CT studies that include the heart, specifically for adult patients over 30 years old. The software employs an artificial intelligence algorithm to automatically detect and estimate coronary artery calcium. This process is integrated into the radiologist's standard workflow, receiving CT studies via Zebra’s Imaging Analytics Platform (IMA), validating them for compatibility, and then performing the analysis. The system is intended to be used opportunistically, leveraging the frequent use of CT scans for other purposes, like lung cancer screenings, to identify coronary calcification. The HealthCCSng algorithm quantifies calcium deposits (identified at 130 Hounsfield Units and above) and generates a three-tier estimated coronary calcium detection category: Low (0-99), Medium (100-399), and High (>400). Alongside this categorical output, it provides preview axial images displaying the detected calcium with annotations, intended for informational purposes to support the radiologist, not for stand-alone diagnosis or clinical decision-making. Performance was validated in a retrospective study using 447 anonymized CT chest cases, comparing its output to a ground truth established by radiologist consensus. The device demonstrated an overall agreement of 92.5% with the ground truth, successfully meeting its performance goals across various imaging parameters and CT scanner manufacturers.

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 HealthCCSng algorithm differentiate coronary artery calcification from other high-density structures or artifacts within the thoracic cavity on non-cardiac-gated CTs, which often have motion artifacts?; What is the specific impact on diagnostic accuracy or patient management when using a 3-category calcium score (Low, Medium, High) from HealthCCSng compared to the more granular Agatston scoring or the 4-category system of its predicate device?; Given its opportunistic nature with non-cardiac-gated CTs, what is the sensitivity and specificity of HealthCCSng for detecting clinically significant coronary artery calcium in populations undergoing CT scans for non-cardiac indications, and how does this vary by the primary reason for the CT scan?; How does the performance of the HealthCCSng algorithm, specifically its 92.5% overall agreement, vary across different CT scanner manufacturers, slice thicknesses, radiation doses, and reconstruction algorithms beyond the general statement of being 'well supported'?; What are the typical processing times for HealthCCSng to analyze a CT study and deliver the calcium detection category and preview images within Zebra’s Imaging Analytics Platform (IMA), and how does this integrate with high-volume radiology practices without causing significant delays?