Viz AAA

Viz AAA is a radiological software tool that employs an artificial intelligence algorithm to analyze CTA images of the abdomen, aiming to identify suspected infra-renal, fusiform abdominal aortic aneurysms of at least 3cm in diameter to assist in workflow triage for vascular specialists. Its methodology involves automated image processing on a backend server and subsequent notification via a mobile application, enhancing the prioritization of potentially critical cases for informational, non-diagnostic review, as supported by clinical data demonstrating high sensitivity and specificity.

Viz AAA is a software device designed for radiological computer-assisted triage and notification, specifically for analyzing Computed Tomography Angiography (CTA) images of the abdomen. Its primary function is to support hospital networks and vascular or endovascular specialists in managing their workflow by identifying and prioritizing patient studies that exhibit suspected abdominal aortic aneurysms (AAAs). The system utilizes an artificial intelligence algorithm, specifically a machine learning model, to automatically analyze CTA scans upon receipt at a backend server. Studies flagged with a suspected AAA are then highlighted within a standalone mobile application, enabling specialists to view these prioritized cases alongside compressed, non-diagnostic preview images, thereby facilitating quicker attention to potentially critical findings in parallel with standard care image interpretation procedures. The Viz AAA software is specifically engineered to detect infra-renal, fusiform abdominal aortic aneurysms that are at least 3 cm in diameter. It is important to understand that the device serves as a triage tool and is not intended for diagnostic use; it does not modify original medical images, and its outputs are purely informational, designed to be used in conjunction with comprehensive clinical information and the professional judgment of specialists. The ultimate diagnostic responsibility remains with the vascular or endovascular specialists who review the original images. Clinical performance data from a study of 466 CTA scans indicated high efficacy, achieving a sensitivity of 96% and a specificity of 95% in detecting AAAs, with an area under the ROC curve (AUC) of 0.99. While the device demonstrated robust performance across various subgroups, its sensitivity was notably lower (0.67) when metallic artifacts obscured the aorta.

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 AI algorithm differentiate between true infra-renal, fusiform AAAs (≥3cm) and other aortic abnormalities, calcifications, or imaging artifacts, particularly in cases with suboptimal image quality not involving metallic implants?; What were the specific demographic and clinical characteristics of the patient populations included in the training and validation datasets for the AI/ML algorithm, and how does the device account for anatomical variations or comorbidities that might mimic AAAs?; Given the significant reduction in sensitivity (0.67) when metallic artifacts obscure the aorta, what specific guidance or warnings are provided to users, and are there plans to improve algorithm robustness in the presence of such artifacts?; How does the Viz AAA system integrate with diverse existing hospital Picture Archiving and Communication Systems (PACS) and radiology information systems (RIS) for seamless data ingestion and workflow integration, and what are the typical deployment and training requirements for end-users (vascular/endovascular specialists)?; Beyond the stated sensitivity and specificity, what is the average notification time from image acquisition to specialist alert, and are there ongoing studies to assess the real-world impact on patient outcomes or time-to-intervention for suspected AAAs?