MeVis Liver Suite

MeVis Liver Suite is an image analysis software intended to support physicians in visualizing and quantitatively evaluating hepatic structures from CT and MR images for applications including treatment planning, preoperative surgical evaluation, and post-procedure assessment. It employs a combination of manual segmentation tools and AI-based semi-automatic pre-segmentation for organs, vascular networks, and lesions, enabling 3D visualization, volumetric calculations, tumor margin analysis, and the estimation of liver-to-body weight ratios, with its performance validated through non-clinical studies.

MeVis Liver Suite is an image analysis software designed for visualizing hepatic imaging studies derived from CT and MR scanning devices. It assists physicians in evaluating the liver, its associated vascular anatomy, and calculating volumes of the liver and its vascular territories, supporting workflows for treatment planning, preoperative surgical strategy evaluation, and post-procedure or therapy follow-up assessment of the hepatic system. The software processes single- and multiframe DICOM images, including multiphase contrast-enhanced CT, dynamic contrast-enhanced MRI, and MRCP, with specific requirements for image homogeneity and resolution. Key manual functionalities include segmentation of abdominal organs (such as liver, spleen, pancreas), liver-related vascular structures (bile ducts, hepatic/portal veins, hepatic artery, IVC), and lesions, alongside user-defined labeling and 3D visualization using tools like freehand contouring and region growing. In addition to manual tools, MeVis Liver Suite offers AI-based semi-automatic pre-segmentation for the liver, hepatic artery, hepatic vein, and portal vein, specifically for CT scans of potential living liver donors with healthy livers using defined contrast phases. Users can evaluate segmented objects by calculating volumes (via voxel-counting) of organs, lesions, and defined liver parts (using separation planes or derived vascular territories). The software also enables 3D visualization of user-defined tumor margins relative to vascular structures and calculates liver volume to body weight ratios from user-provided data. It supports manual spatial registration for comparing segmentations across different image series and provides reporting in DICOM and HTML formats. Performance validation included achieving a median DICE score of 0.98 for AI liver segmentation and positive expert assessment for AI vascular segmentation accuracy, with volume calculations verified against phantom and clinical 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 does the performance of the AI-based semi-automatic pre-segmentation tools vary with different CT scanner manufacturers, acquisition protocols, or in livers with subtle pathologies not meeting the strict 'healthy donor' criteria?; What is the quantitative impact on surgical planning time or inter-observer variability in liver segment volumetry when using MeVis Liver Suite's outputs compared to traditional manual methods alone?; Beyond DICE scores and Hausdorff distance for AI liver segmentation, what specific anatomical regions or vascular branch orders present the most challenge for the semi-automatic segmentation tools, and how does required user correction time correlate with these challenges?; How does the accuracy of manual spatial registration, particularly between CT and MRI datasets with inherent differences in resolution and patient positioning, affect the reliability of combined volumetric assessments or tumor margin analysis involving structures segmented from different modalities?; What is the typical learning curve for radiologists and surgeons to achieve proficiency with the full suite of MeVis Liver Suite's manual and AI-assisted tools, especially for complex cases involving atypical vascular anatomy or multiple, closely spaced lesions?