CardIQ Suite

CardIQ Suite is a non-invasive software application designed to enhance the diagnosis and management of cardiovascular diseases by analyzing CT cardiac DICOM data. It employs advanced image processing, including new deep learning algorithms for automated segmentation, labeling, and quantification of cardiac structures and calcifications, thereby aiming to improve diagnostic accuracy and workflow efficiency for clinicians.

CardIQ Suite is a non-invasive software application designed to analyze cardiovascular anatomy and pathology using 2D or 3D CT cardiac DICOM data, including both non-contrast and angiography images. Its primary purpose is to aid clinicians in the diagnosis and determination of treatment paths for various cardiovascular diseases, such as coronary artery disease, by providing tools for the visualization and measurement of vessels, assessment of chamber mobility, and follow-up for interventions like stent placement and bypasses. The software offers comprehensive calcium scoring capabilities, employing a deep learning algorithm for automatic segmentation and labeling of calcified plaques in coronary arteries, heart valves, and the aorta, utilizing standard Agatston/Janowitz and Volume scoring methods. It also allows for manual lesion scoring and editing, with results exportable in DICOM SR format. The CardIQ Suite leverages advanced image processing, centrally featuring new deep learning algorithms to enhance workflow efficiency for several critical functions. These AI-driven enhancements include automated segmentation of the heart for 3D volume rendering, automatic segmentation of the coronary tree, precise tracking of coronary centerlines, and accurate labeling of coronary arteries, functionalities which were previously manual or employed traditional algorithms in earlier versions or reference devices. The software also provides Multi-Planar Reformat (MPR) Cardiac Review tools with cine capabilities for dynamic visualization of cardiac structures. Non-clinical bench testing successfully validated the performance of these four new deep learning algorithms. Furthermore, a clinical reader study confirmed their practical utility, demonstrating that the automated outputs for heart and coronary analysis were deemed acceptable in over 90% of evaluated cases with good image quality, supporting the claim of improved workflow efficiency.

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 (accuracy, processing time) of the new deep learning algorithms for heart and coronary segmentation compare quantitatively to the traditional algorithms used in the reference device (K073138) across diverse patient populations and image qualities?; What specific training datasets and validation methodologies were employed for the four new deep learning algorithms, and what is their robustness against variations in CT acquisition protocols (e.g., scanner types, contrast timing, radiation dose)?; Beyond workflow efficiency, does the enhanced automated segmentation and labeling provided by CardIQ Suite lead to improved inter-observer and intra-observer variability in measurements and diagnostic interpretations compared to manual or semi-automated methods?; What are the specific limitations of the deep learning algorithms, particularly in cases of atypical anatomy, severe artifact presence, or complex pathologies, and what are the recommended user interventions or manual override procedures in such scenarios?; How does the "Adaptive Volume scoring" method for calcium scoring differ from the standard Volume scoring, and what is the clinical evidence supporting its diagnostic or prognostic value?