Medical Image Post-processing Software (uOmnispace.CT)

uOmnispace.CT is post-processing software for viewing, manipulating, evaluating, and analyzing medical images, primarily CT scans. It contains 15 applications for specialized analysis including lung density, vessel analysis, cardiac imaging, liver evaluation, brain perfusion, and body perfusion, supporting interpretation and clinical decision-making in healthcare institutions.

The proposed device introduces deep-learning algorithms for segmentation and extraction tasks (lung, airway, bone removal, coronary artery, heart chamber, liver, and hepatic vessels) and integrates both rigid and non-rigid motion correction. These enhancements provide improved automation while maintaining the same fundamental post-processing and visualization framework as the predicate device.

NEMA PS 3.1–3.20 (DICOM), ISO 14971 (risk management), and IEC 62304 (software life cycle processes). Software verification and validation included hazard analysis, requirements specification, architecture design, and cybersecurity documentation. ML/AI algorithms were validated using Dice similarity coefficient against ground truth annotations across diverse patient populations.

The proposed device performs substantially the same intended functions as the predicate (K233209) with identical classification, regulatory status, and clinical use. Although deep-learning algorithms enhance automation and motion correction methods improve accuracy, these represent iterative algorithmic improvements that do not alter the fundamental post-processing technology or clinical purpose. Performance testing demonstrates the algorithms meet or exceed acceptance criteria across diverse patient subgroups, maintaining safety and effectiveness equivalent to the predicate device.