Neosoma Brain Mets

Neosoma Brain Mets is software designed to automatically segment (contour) known or previously diagnosed brain metastases on MRI images to assist qualified medical professionals. It uses artificial intelligence and deep learning to generate initial Gross Tumor Volume contours that clinicians must review and finalize before clinical use. The software does not alter original images, is not for tumor detection or diagnosis, and is intended for adult patients only.

The device uses machine learning-based semi-automatic segmentation of brain metastases from T1 post-contrast MRI sequences, with volumetric measurements and color-coded overlays as output. It leverages the same pre and post-processing pipeline components (MRI normalization, atlas registration, skull stripping) as the reference device NS-HGlio. The architecture provides 2D and 3D image review capabilities with non-machine-learning post-processing for volume calculations.

Software validation followed IEC 62304:2006/AC:2015 (Medical device software – Software life cycle processes) and the 2023 FDA Guidance document 'Content of Premarket Submissions for Device Software Functions.' Clinical performance was evaluated using standard quantitative endpoints including sensitivity, false positive rate, Dice Similarity Coefficient (DSC), Hausdorff Distance (HD95), and Mean Surface Distance (MSD).

Neosoma Brain Mets is substantially equivalent to predicate device VBrain because both use identical AI/deep learning algorithms for semi-automatic segmentation of known brain tumors on T1 post-contrast MRI, generate GTV contours for informational purposes only, do not replace manual contouring, do not alter original images, and are not intended for diagnosis. Both target the same anatomical site (brain) and clinical workflow (radiation oncology planning). The subject device meets or exceeds all performance acceptance criteria across diverse patient populations and imaging equipment.