Auto-Seg (SO-0012), Spine Auto-Seg (SO-0012)
Agada Medical, Ltd.Auto-Seg (SO-0012) by Agada Medical, Ltd. is software that processes CT images of the spine and pelvis to automatically segment and label vertebrae and pelvis bones. It assists spine surgeons by providing detailed 3D segmentation outputs that aid in the assessment of musculoskeletal diseases and surgical planning for adult patients. The software uses validated AI algorithms and does not provide direct diagnoses but supports clinical decision-making.
Intended Use
Image processing software to provide qualitative analysis of CT DICOM images of spine and pelvis to support spine surgeons in assessment and surgical planning for adults aged 18 and older. Performs segmentation and labelling of vertebrae and pelvis objects but does not provide diagnosis or treatment recommendations.
Technology
Software-only AI/ML image processing system using four convolutional neural network models to segment and label spine (thoracic and lumbar vertebrae) and pelvis in CT images. Algorithms output 3D segmentation masks in DICOM format, STL, and NIfTI files. Uses non-adaptive machine learning models with graphical user interface for workflow including user confirmation of vertebra labeling. Runs on standard personal computer with GPU. Developed with cybersecurity controls per FDA guidance.
Performance
Device underwent extensive non-clinical validation including software verification/validation, labeling verification, risk management, human factors/usability testing, cybersecurity testing, and standalone performance assessments using publicly available datasets. Performance was assessed via segmentation accuracy (DICE coefficient), labeling accuracy, and processing time, all meeting acceptance criteria with high accuracy (>99%), segmentation DICE >0.8 for spine and pelvis, and segmentation time within 10 minutes. Confirmatory non-inferiority testing using independent clinical sites also demonstrated robust and generalizable performance.
Predicate Devices
No predicate devices specified
Ready to Sharpen Your Edge?
Subscribe to join 11k+ peers who rely on RadAI Slice. Get the essential weekly briefing that empowers you to navigate the future of radiology.
We respect your privacy. Unsubscribe at any time.