AVIEW

AVIEW is an advanced image processing software that leverages deep-learning algorithms to provide quantitative analysis of CT thoracic and cardiac datasets, aiding physicians in diagnosing and managing pulmonary diseases and assessing cardiovascular risk. Its methodology encompasses automated segmentation of anatomical structures, detailed characterization of lung nodules including Lung-RADS classification, and coronary artery calcification analysis with Agatston scoring.

AVIEW is an image processing software designed for the review and analysis of CT thoracic and cardiac datasets. Its purpose is to support physicians quantitatively in diagnosis, follow-up evaluation, and documentation by providing image segmentation of sub-structures in the lung (lobes, airways) and heart, registration of inspiration/expiration CT scans for analyzing air trapping volume and inspiration/expiration ratio, and tools for volumetric/structural analysis, density evaluation, and reporting. The software, installable on a PC and accessible on-premise or via cloud (including mobile devices), utilizes DICOM 3.0 and employs deep-learning algorithms for automatic lung/lobe/airway segmentation and inspiration/expiration registration. The software offers comprehensive characterization of lung nodules, including type, location, size, volume, Hounsfield Unit (HU) values, mass, volumetric measures, Volume Doubling Time (VDT), and Lung-RADS classification, integrating with third-party CAD for nodule detection. For cardiac assessment, AVIEW performs Coronary Artery Calcification (CAC) analysis by segmenting four main arteries, extracting calcium, providing Agatston, volume, and mass scores, and assessing CAC risk based on age and gender, utilizing deep learning for calcium segmentation. Additional features include COPD analysis, LAA analysis with 3D cluster sizing, airway wall thickness measurement, a 'Lobar-pealing method' for pulmonary vessel analysis, and export of over 1000 feature values for radiomics research. Its performance and functionality were established through non-clinical software verification and validation testing.

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. What is the documented accuracy and processing speed of the deep-learning algorithms for lung, lobe, and airway segmentation compared to manual segmentation or other AI tools?; How does AVIEW's 'Lobar-pealing method' for pulmonary vessel analysis technically operate, and what are its specific advantages in clinical scenarios over conventional vessel analysis techniques?; What validation studies support the accuracy of the Coronary Artery Calcification (CAC) risk assessment based on age and gender, particularly across diverse patient populations?; Can the software differentiate and quantify various components of complex lung nodules (e.g., solid vs. ground-glass opacity proportions), and how does this influence Lung-RADS categorization and Volume Doubling Time (VDT) calculation?; What are the specific '1000+ feature values' exported for radiomics research, and how are the segmentation masks stored in open formats (Analyze or NifTi) structured for such research purposes?