AIBOLIT 3D+

Aibolit 3D+ is a medical image management and processing system designed to enhance surgical planning and intraoperative decision-making by converting CT and MRI DICOM images into interactive 3D anatomical models. Its methodology relies on sophisticated image processing, including radiologist-controlled, AI-assisted organ segmentation and 3D reconstruction, validated for accuracy in representing anatomical structures.

Aibolit 3D+ is a software-based medical imaging system designed to assist healthcare professionals in preoperative surgical planning, training, patient education, and intraoperative guidance. It processes multi-dimensional digital images acquired from CT and, critically with this update, MR imaging devices, converting DICOM files into interactive 2D and 3D anatomical models. The system allows users to manipulate these models, offering features like multi-axis rotation, transparent viewing of internal structures such as ducts, vessels, lesions, and calcifications, and differential coloration for highlighting specific anatomical regions. The workflow involves physicians uploading images and specifying areas of interest, followed by radiologist-led annotation and segmentation, optionally assisted by AI/machine learning, before the 3D model is generated for review and use. The core science involves advanced image processing algorithms for organ segmentation and 3D model reconstruction. For image segmentation, Aibolit 3D+ incorporates an optional artificial intelligence component based on machine learning, trained on a dataset of 108 anatomical structures from MRI scans; this AI-assisted segmentation is reviewed and controlled by a radiologist. Performance testing, including software verification and validation, demonstrated the system's intended use. Specific tests validated the segmentation software, assessed the repeatability and reproducibility of segmenting various anatomical structures, and confirmed the accuracy of 3D model generation. Further validation, particularly for the newly added MRI image processing capabilities, confirmed the conservation of shape dimensions and volume against a "ground truth" standard across multiple organ structures, radiologists, and viewing perspectives.

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 AI-assisted segmentation algorithm's performance compare to manual segmentation by experienced radiologists in terms of accuracy and time efficiency for complex anatomical structures derived from MRI?; What is the specific "ground truth" standard used for validating the conservation of shape dimensions and volume, and how was its accuracy established?; What are the limitations of the software in processing images with significant artifacts or from patients with highly atypical anatomy?; How does the system ensure data integrity and security, particularly when handling sensitive patient DICOM data through its web-based platform and during AI processing?; What quantitative metrics were used to assess the accuracy of the 3D model generation for various organ structures, and what were the acceptable thresholds for these metrics?