Arterys Oncology DL

Arterys Oncology DL is an advanced medical imaging software designed to augment oncological review by providing tools for multi-modal (CT/MR) image visualization, deep learning-assisted semi-automatic segmentation of lung and liver lesions, and quantitative analysis. Its purpose is to support clinicians in efficiently evaluating tumor burden, tracking changes over time through longitudinal analysis, and standardizing documentation with Lung-RADS/LI-RADS reporting, thereby complementing traditional diagnostic workflows.

Arterys Oncology DL is a medical diagnostic software application designed to assist clinicians in the oncological workflow. It facilitates the viewing, manipulation, 3D-visualization, and comparison of medical images, primarily CT and MR scans, from multiple imaging studies and across different time points. Key functionalities include 2D and 3D image review, manual volumetric segmentation, and notably, semi-automatic volumetric segmentation of lung nodules from CT scans and liver lesions from MR scans, leveraging deep learning models. The software also offers co-registration of images, longitudinal tracking of lesions, and quantification of nodule/lesion size and volume, aiming to streamline the assessment, monitoring, and documentation of lesions. The system utilizes advanced image processing techniques, particularly its deep learning algorithms for targeted semi-automatic segmentation tasks, and supports standardized reporting based on guidelines like Lung-RADS and LI-RADS, aiding in consistent documentation and the assessment of treatment response. Arterys Oncology DL is intended as a complementary tool to existing diagnostic procedures, offering enhanced visualization and analytical capabilities to help clinicians assess morphological changes during diagnosis and therapy follow-up. While providing sophisticated analytical tools, the software underscores that the ultimate responsibility for making pertinent diagnoses remains with the clinician, who should use it in conjunction with their standard practices and visual comparison of the original image data.

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 robust is the semi-automatic volumetric segmentation for atypical lung nodules (e.g., subsolid, ground-glass) or complex liver lesions (e.g., post-treatment, infiltrative), and what is its performance benchmark against expert manual segmentation in diverse patient populations and varying image qualities?; What specific deep learning architectures are employed for the lung CT nodule and liver MR lesion segmentation, and how were these models trained and validated (including datasets used) to ensure generalizability across different scanner vendors, acquisition parameters, and contrast phases?; Can the longitudinal tracking feature accurately account for significant anatomical changes or deformations between imaging time-points, particularly in organs susceptible to respiratory motion, tumor response heterogeneity, or treatment-induced alterations like fibrosis?; What is the quantitative impact of using Arterys Oncology DL on inter-observer and intra-observer variability in lesion measurement (e.g., RECIST, mRECIST) and characterization, and how does it affect diagnostic workflow efficiency (e.g., reading time) compared to traditional PACS viewing and manual methods?; How does the co-registration algorithm handle images from different modalities (e.g., CT and MR, or pre- and post-contrast CT) with varying resolutions, slice thicknesses, and fields of view, and what is its validated spatial accuracy for aligning lesions across these datasets for comparative assessment, especially for smaller lesions?