AVIEW LCS

AVIEW LCS is an image processing software designed to aid in the characterization of lung nodules from thoracic CT images using semi-automatic segmentation, comprehensive measurements, and CAD integration. It leverages deep-learning for anatomical segmentation, applies the PANCAN risk model for malignancy assessment, and uses Lung-RADS for standardized reporting and tracking nodule evolution with VDT.

AVIEW LCS is an advanced image processing software designed for the detailed review, analysis, and reporting of thoracic CT images, specifically focusing on the characterization of lung nodules. Its core purpose is to assist clinicians by providing comprehensive tools for nodule assessment, whether in a single study or across multiple studies over time for follow-up. The system facilitates semi-automatic nodule segmentation and offers a suite of measurements including nodule type, precise location, size (major/minor axes, effective diameter from volume), overall volume, Hounsfield Unit (HU) statistics (mean, min, max), mass, and specific volumetric details of solid nodule components. It also supports workflow integration by allowing the import of results from compatible Computer-Aided Detection (CAD) systems for automated nodule detection. The scientific methodology underpinning AVIEW LCS incorporates several advanced features. It utilizes deep-learning algorithms for improved segmentation of lungs and lobes, enhancing the accuracy of nodule localization and analysis. For risk stratification, the software calculates a malignancy score based on the PANCAN risk model. AVIEW LCS automates the categorization of findings using the Lung-RADS classification system, promoting standardized reporting and management recommendations. Furthermore, it supports longitudinal tracking through automated nodule matching between scans and calculates Volume Doubling Time (VDT) to assess nodule growth. The system generates comprehensive reports, including PDF and DICOM structured reports, encapsulating all findings and classifications.

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 are the specific performance metrics (e.g., Dice similarity coefficient) for the deep-learning based lung and lobe segmentation algorithms, especially in challenging cases with adhesions or severe pathology?; How does the AVIEW LCS system handle variations in CT acquisition parameters (e.g., slice thickness, reconstruction kernel) when performing nodule measurements and calculating VDT across follow-up scans?; What validation studies were performed to assess the accuracy and reliability of the PANCAN risk model implementation within AVIEW LCS across diverse patient cohorts and nodule characteristics?; How does the semi-automatic nodule segmentation tool perform with various nodule morphologies (e.g., spiculated, lobulated, ground-glass, part-solid), and what is the typical user interaction time required for accurate segmentation?; Regarding the Mevis CAD integration, how are discrepancies between CAD-detected nodules and user-identified nodules managed within the AVIEW LCS workflow, and how does the system ensure all relevant findings are incorporated into the final report?