Automated Aortic Stenosis Software (AutoAS)

AutoAS is a software application that uses artificial intelligence to analyze transthoracic echocardiography images and assist physicians in assessing moderate to severe aortic stenosis. It provides a binary classification (suggestive or not suggestive of moderate/severe AS) with an associated confidence metric, intended as a diagnostic aid to determine if further evaluation is needed. The software is not intended for standalone clinical decision-making and must be reviewed by a qualified physician.

AutoAS is a machine learning-based software that processes B-mode transthoracic cardiac ultrasound images from specific views (PLAX, PSAX-AV, AP5) to generate a two-level binary output with a study-level confidence score. It incorporates a clip annotator to verify valid image views and includes heart rate estimation for internal use. The device is designed as an accessory to compatible GE HealthCare ultrasound systems.

Not stated in this summary. The document references FDA's Guidance 'Content of Premarket Submission for Device Software Functions' (June 14, 2023) for software verification and validation, and compliance with GE HealthCare's Quality Management System and ISO 13485 standards.

AutoAS is substantially equivalent to EchoGo Pro (K201555) because both are machine learning-based diagnostic aids for cardiac ultrasound assessment with similar intended use, identical classification (892.2060), identical product code (POK), and identical regulation (21 CFR 892.2060). Although AutoAS assesses aortic stenosis while EchoGo Pro assesses coronary artery disease, both operate within the same anatomical area and are designed to support physician decision-making without replacing clinical judgment. Extensive performance testing demonstrated AutoAS achieves comparable or superior performance metrics (93.2% AUC, 92.4% specificity, 75.2% sensitivity) relative to expert cardiologists.