AmCAD-UT

AmCAD-UT is a software device intended to aid medical professionals in analyzing thyroid ultrasound images by offering AI-driven or user-defined region of interest contouring for nodules larger than 1cm. It employs AI, statistical pattern recognition, and quantification methods to provide detailed visualization and measurement of sonographic characteristics, thereby supporting clinical assessment for potential biopsy.

AmCAD-UT is a Windows-based computer-aided detection (CADe) software application developed by AmCad BioMed Corporation. It is designed to assist medical professionals in the analysis of thyroid ultrasound images obtained from standard ultrasound systems. The primary function of AmCAD-UT is to facilitate the detailed examination of user-selected thyroid nodules, specifically those larger than 1cm for which a biopsy recommendation is being considered. The system allows for the region of interest (ROI) around a nodule to be manually defined by the user or automatically suggested by an integrated AI contouring algorithm after the user indicates the nodule's location. Once the ROI is established and confirmed, AmCAD-UT processes the image to provide quantification and visualization of various sonographic characteristics. These characteristics include hyperechoic foci, echogenicity, texture, margin, orientation, and the presence of anechoic areas. The underlying technology combines artificial intelligence, statistical pattern recognition, and quantification methods to analyze these features. The software automatically generates a report based on user inputs and the analytical findings, which can be printed or saved as a PDF. Standalone performance studies have been conducted to validate the effectiveness of the AI-suggested ROI contouring, using ROIs labeled by a panel of specialists as the ground truth, demonstrating its utility in accurately determining nodule contours.

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 of the AI-suggested ROI contouring compare to manual ROI definition by experienced radiologists in terms of accuracy, time-efficiency, and inter-observer variability for nodules larger than 1cm?; What specific AI algorithms and statistical pattern recognition techniques are employed for the quantification and visualization of sonographic characteristics (e.g., hyperechoic foci, echogenicity, texture, margin), and how were these models trained and validated against ground truth?; Are there particular types or complex presentations of thyroid nodules (e.g., those with poorly defined margins or cystic components) where the AI contouring algorithm or sonographic characterization shows limitations?; How does the integration of AmCAD-UT's quantification and visualization features influence a physician's diagnostic confidence and subsequent biopsy recommendations for thyroid nodules compared to standard ultrasound review alone?; What is the reported accuracy or performance metrics (e.g., Dice similarity coefficient for ROI, sensitivity/specificity for characterization) of the AI-suggested contouring and sonographic feature analysis based on the standalone performance studies using specialist-labeled ROIs as ground truth?