GBrain MRI

GBrain MRI is post-processing medical device software that analyzes brain MRI images to automatically segment and quantify signal hyperintensities on T2-weighted FLAIR and T1-weighted post-contrast images. It provides volumetric measurements and visual overlays to aid radiologists in quantitative assessment of structural brain MRIs in diagnostic radiology contexts, while emphasizing that physician review and clinical judgment remain essential.

Both the subject and predicate devices are software-only systems that accept DICOM-format brain MRI images as input and output volumetric measurements in DICOM format. Both use deep learning algorithms to segment hyperintensities, calculate volumes, generate colored segmentation overlays on DICOM images, and summarize results in PDF reports. The subject device extends the predicate's functionality by adding segmentation capability for contrast-enhancing regions on T1-weighted post-contrast images in addition to FLAIR images.

Not stated in this summary.

The subject device shares identical intended use, intended users, environment of use, patient population, and regulatory classification (21 CFR 892.2050, Class II) as the predicate. Both perform the same core technical task—automated quantification of MR signal hyperintensity for diagnostic support—using equivalent deep-learning-based segmentation methodology. Performance testing on 131 validation cases across multiple MRI manufacturers and field strengths demonstrated that contrast enhancement measurements achieved R² ≥0.94 and DICE ≥0.81, meeting planned acceptance criteria and showing similar accuracy to the predicate. The addition of T1-weighted post-contrast image analysis is a natural extension of the same segmentation principle already proven in the predicate's FLAIR processing, raising no new safety or effectiveness concerns.