Arterys MICA

Arterys MICA is a cloud-hosted medical image management and processing system designed to assist healthcare professionals in diagnosis by enabling the display, manipulation, quantitative analysis of DICOM and non-DICOM medical images, and the display of 3rd party AI model outputs. It employs a modular architecture with optional AI-driven modules for cardiac (including T1/T2 mapping) and oncological applications, utilizing deep learning for tasks like segmentation and parametric mapping to enhance diagnostic workflows.

Arterys MICA is a cloud-hosted medical image processing software designed for viewing, managing, and analyzing DICOM and non-DICOM medical images, such as CT or MR scans. Accessed via a web browser, it serves as a support tool for clinicians by providing capabilities for image storage, filtering, digital manipulation, and quantitative measurements. The core Viewer AI component offers a comprehensive image viewing environment with features like 2D/3D/MIP display, study list management, view synchronization, metadata access, and various measurement tools. An enhancement allows Viewer AI to display outputs from third-party AI models, including segmentations, measurements, bounding boxes, and heatmaps, broadening its analytical display capabilities. The system's functionality is extended by optional modules: Cardio AI and Oncology AI, both leveraging deep learning algorithms to automate time-consuming manual tasks like segmentation and landmark identification, with results remaining user-reviewable and editable. The Cardio AI module specializes in cardiovascular MR image analysis, quantifying blood flow, ventricular function, analyzing perfusion and delayed enhancement datasets, and performing parametric mapping (T1, T2, T2* values, ECV calculations) to assess myocardial tissue changes; recent additions include cardiac MRI T1 and T2 workflows. Performance validation for these T1/T2 workflows showed values within predefined accuracy limits (30 ms for T1, 4 ms for T2) compared to OEM vendor maps, with deep learning-generated contours for T1 mapping performing as intended. The Oncology AI module provides tools to assess and document morphological changes in lesions for diagnostic and therapy follow-up.

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 specific impact of the cloud-hosted architecture (AWS) on data security, processing speed, and accessibility for users in different geographical locations with varying internet infrastructure?; How does the performance of Arterys MICA's deep learning algorithms for segmentation and landmark identification (e.g., in Cardio and Oncology AI) compare to other AI tools or manual methods in terms of accuracy, consistency, and time-saving across diverse patient populations and image qualities?; For the Cardio AI module's T1 and T2 mapping, what is the range of variability attributable to user-dependent ROI selection, and what best practices are recommended to ensure inter-observer and intra-observer reproducibility for these quantitative measures?; Regarding the Oncology AI module, what specific types of lesions and morphological changes has it been validated to assess, and what is its performance in tracking subtle changes over time compared to standard radiological assessment protocols (e.g., RECIST criteria)?; With the Viewer AI's capability to display outputs from 3rd party AI models, what mechanisms are in place to ensure clear differentiation between Arterys MICA's native analyses and external model outputs, and what are the technical requirements for integrating these third-party models effectively?